Rooftop Revolution: Unleashing Delhi's Solar Potential

© 2013 GREENPEACE India List of Figures 7 For further enquiries, Research/Authors: List of Tables 8 please contact: Tobias Engelmeier, BRIDGE TO INDIA Preface 10 Mohit Anand, BRIDGE TO INDIA Mr. Jasmeet Khurana Jasmeet Khurana, BRIDGE TO INDIA 1 executive Summary 12 Jasmeet.khurana@ Prateek Goel, BRIDGE TO INDIA 2 Introduction 15 Tanya Loond, BRIDGE TO INDIA bridgetoindia.com 2.1 Solar in India 15 2.2 Power in Delhi and why solar makes sense 17 Mr. Abhishek Pratap Concept: [email protected] Manish Ram, GREENPEACE 3 Solar resource availability for Delhi 21 Abhishek Pratap, GREENPEACE 4 Delhi’s geographic potential for solar rooftop installations 23 Anand Prabhu Pathanjali, GREENPEACE 4.1 Residential buildings 26 4.2 Commercial buildings 26 Report produced by: 4.3 Industrial buildings 26 GREENPEACE India 4.4 Government buildings 27 June 2013 4.5 Public and semi-public facilities 27 4.6 Transport facilities 27 Design: tiffinbox.in 5 Delhi’s existing solar policies and incentives 28 6 Integration of solar PV with the grid 29 Printed on 100% recycled paper 7 the viability of rooftop solar PV 32 7.1 Viability of rooftop solar PV for commercial consumers 38 7.2 Viability of rooftop solar PV for industrial consumers 39 7.3 Viability of rooftop solar PV for residential consumers 40 7.4 Viability of rooftop solar PV for government facilities 42 8 Business models for rooftop solar 44 8.1 Straight-forward sales model 44 8.2 Renewable Energy Service Company (RESCO) Model 45 8.3 Local Micro Utility Model 46 9 why rooftop solar works for the stakeholders 50 9.1 Why rooftop solar works for Delhi’s power consumers 50 9.2 Why rooftop solar works for Delhi’s DISCOMS 51 9.3 Why rooftop solar works for Delhi’s government 53 10 Delhi solar roadmap: reaching 2 GW by 2020 56 10.1 Priority scenario (parity based) 56 10.2 Optional scenario (government incentives) 58 11 government policy recommendations 62 11.1 Recommendations for enabling parity based solar growth 62 11.2 Recommendations for activating early adoption 70 12 APPENDIX 76 12.1 Methodology for calculating the rooftop solar potential in Delhi 76 12.2 Computation of the solar suitable rooftop space for 77 different land areas 12.2.1 Residential buildings 77 12.2.2 Commercial buildings 81 12.2.3 Industrial Buildings 82 12.2.4 Government buildings 83 12.2.5 Public and semi-public facilities 83 12.2.6 Transport buildings 84 12.2.7 Factors that could impact the potential for rooftop solar 85 List of Figures 12.3 Ways to increase Delhi’s geographic potential for solar 86 12.4 Rooftop solar PV systems 86 Figure 1: State-wise commercial tariff vs. LCOE of solar power (100 kWp 12.5 Solar PV grid connectivity challenges and solutions 92 system without battery) and diesel generation (100 kVA system) Figure 2: India’s structural power deficit 13 glossary of Terms 94 Figure 3: India’s rising power costs (2000-2018) Figure 4: Irradiation maps of Delhi, Beijing and Munich Figure 5: Map of Delhi (NCT) Figure 6: Deriving Delhi’s rooftop space for solar Figure 7: Potential of rooftop solar power generation in Delhi for different land area types (MW) Figure 8: Centralized vs mixed (centralized and distributed) power generation Figure 9: Historical grid tariff price trend Figure 10: Falling prices of solar PV modules over the past 32 years Figure 11: System prices have fallen over 47% in last 10 years Figure 12: Solar costs in India in 2013 for different system sizes (without storage) Figure 13: Delhi rooftop consumers’ market size and solar viability Figure 14: Viable solar tariff for commercial consumers vs. non-domestic grid tariff (type-1 and type-2) Figure 15: Viable solar tariff for industrial consumers vs. industrial grid tariff (type-1 and type-2) Figure 16: Residential solar investment compared to similar low-risk investment options Figure 17: Payback from investment in a 5 kW residential system Figure 18: Viable solar tariff for residential consumers vs. domestic grid tariff Figure 19: Viable solar tariff for government buildings vs. domestic grid tariff Figure 20: The implementation structure of the Gandhinagar Solar Rooftop Program Figure 21: Gross metering Figure 22: Benefits to consumers Figure 23: Benefits to the DISCOMS Figure 24: Benefits to the government Figure 25: Priority scenario expected annual solar capacity additions (in MW) Figure 26: Priority scenario cumulative solar rooftop growth until 2020 Figure 27: Optional scenario expected solar capacity addition (in MW) Figure 28: Optional scenario cumulative solar rooftop growth until 2020 Figure 29: Grid handling capacity vs. expected solar growth Figure 30: Policy recommendations for the government Figure 31: Screenshots of the New York solar map Figure 32: Electricity generation and CUF of PV plant at the German House, New Delhi Figure 33: Cost to the government for viability nudges Figure A1: Solar PV system Figure A2: Delhi residential demand compared with solar generation

7 List of Tables

Table 1: Delhi’s solar RPOs and required solar capacity (per FY) Table 2: Annual irradiation data for Delhi Table 3: Solar suitable rooftop area available in different land area types Table 4: Consumer categories for rooftop solar PV Table 5: Viability analysis of rooftop solar PV in Delhi in 2013 Table 6: Factors impacting rooftop solar adoption Table 7: German House solar project details Table A1: Solar suitable areas by jurisdictions Table A2: Assessment of colonies under the jurisdiction of the MCD

Photograph by Anand Pathanjali/Greenpeace India

8 9 While clear targets on renewable energy at the national level would certainly Preface help build investors’ confidence and create a mandatory domestic market, a realistic and rationalized differential Renewable Purchase Obligation Delhi, the world’s greenest capital? (RPO) would help in creating energy equity in the country’s energy delivery system. Though effective implementation of the RPO mechanism in each “Rooftop Revolution – Unleashing Delhi’s Solar Potential” is an in-depth state is essential to ensure renewable energy projects are actually producing analysis of the potential for rooftop solar in Delhi, a city that has often claimed electricity and are not just tax evasion models, there is also a need to revise to be the world’s greenest capital. The report comes at a time when India’s each state’s RPO target with a more rational approach so that intra-country We hope that political nerve-centre and national capital is facing a potential electricity energy equity will be established in line with India’s professed international this report will provide crisis. Demand in the city is expected to reach an all-time high of 6,000 MW position of “common but differential responsibility.” In this regard, Delhi, the the city’s stakeholders this summer while more than 17 states, many of which the city relies on for its second richest state and third highest per capita energy consumer, can set a power, will face significant power shortages. With the country’s eyes on Delhi precedent by initiating a rooftop solar revolution by 2020 and usher in a new with the insights and to be India’s answer to the world’s sustainable city concept, the crippling power era of a democratic energy delivery system that runs on equity, justice and data needed to ensure supply situation reminds us that there is an urgent need to rethink electricity human aspiration. that Delhi truly lives up generation and supply in the city. to its title of the world’s The enthusiasm to support solar in Delhi exists, and through our campaign greenest capital. Within the last decade, Delhi’s electricity demand rose by an average 6% every we will ensure that it will continue to grow. What was still lacking is a Delhi year. From 20 billion units in 2002, the demand in all likelihood will reach over specific, in-depth assessment of the opportunities and challenges for the 33 billion units by 2017, a 65% growth. This poses a fundamental question effective adoption of rooftop solar. The first step towards positioning rooftop of how to ensure a consistent electricity supply. The city’s many shopping solar energy as a solution is to understand its potential in the city. We have malls, Delhi Metro and state-of-the-art health facilities currently get three- initiated this report written by BRIDGE TO INDIA to serve that purpose. We fourth of their electricity supply from outside the state, relying mostly on hope that it will provide the city’s stakeholders with the insights and data environmentally unsustainable fuels like coal. A shift from centralized, fossil needed to ensure that Delhi truly lives up to its title of the world’s From 20 billion units fuel based power sources to decentralized, renewable sources like solaris greenest capital. in 2002, the demand in all then the best way to overcome these hurdles, while achieving other substantial likelihood will reach over benefits like the creation of jobs in the green sector and independence from other states for energy supply. 33 billion units by 2017, a 65% growth. Delhi can “Switch on the Sun” Abhishek Pratap Senior Campaigner As the capital and second richest state in the country, Delhi is in a very good position to take the lead in transitioning to a decentralized, sustainable Greenpeace India energy paradigm. With this conviction, we have launched the “Switch on the Sun”campaign. Our goal is to emphasize the effectiveness of solar energy as a solution to Delhi’s emerging power crisis. In the coming days, the campaign seeks to bring together all the stakeholders in the city: distribution companies, As the capital and government decision makers, regulatory bodies and electricity consumers. second richest state in the country, Delhi is in a The campaign has so far received an excellent response. Through our interactions with resident welfare associations (RWAs), we have had individual very good position to take colonies like Sukhdev Vihar in Delhi pledge their commitment to rooftop solar. the lead in transitioning The power minister of the city has agreed with our message and initiated to a decentralized, government administrators to move ahead on solar with a sense of urgency. sustainable energy It is also encouraging to see that national level bodies like the Central paradigm. Electricity Authority are already drafting regulations that will bring responsibility-focused equitable delivery of energy through decentralized renewable energy systems. We are confident that the Delhi Electricity Regulatory Commission will take the cue from our campaign and initiate the required regulatory changes outlined here at the city level to ensure renewable energy supply reaches its desired level.

10 11 only after 2020. In our roadmap, we suggest that 2 GW of solar power on the 1 eXecutive Summary grid would be reached by 2020. Solar power would not exceed 20% of the load expected on the grid by 2020. So, storage is perhaps not necessary, from a grid In this report, we propose that Delhi can be a 2 GW solar city by 2020. This perspective. However, more research should be conducted on understanding may sound bold when compared to the reality of existing solar installation in the effect of large-scale solar deployment on the grid. It is also reasonable to leading cities such as Berlin (98 MW1), New York (14 MW2) or San Francisco (23 assume that over the next years, new technologies and the experiences of grid MW3) or the overall achievement of 1.5 GW in India as a whole until mid-2013. operators across the world will provide new solutions. However, it is not so bold, if we assume that the landscape of power supply Solar, as a is changing fundamentally: Grid power prices continue to rise and the power Economic and demographic growth in India’s capital Delhi has been locally available, supply deficit in India is widening. Given the dramatic fall in the cost of solar accompanied by a significant expansion of the city’s infrastructure. The Delhi environmentally friendly power (by as much as 50% in the last two years), solar has moved into the government has been committed to improving energy supply. In recent years, and increasingly viable mainstream and become a viable option. Theoretically, the total land area on Delhi has achieved a very high availability of power. Blackouts have become which Delhi is built could support 123 GW of solar PV. Therefore, 2 GW requires source of power can rare. However, given India’s growing power deficit, the rising cost of power only 1.6% of the city’s land. provide Delhi with an and Delhi’s rapidly growing power demand4, it will be difficult to maintain this level of supply stability in the years to come. Solar, as a locally available, attractive long-term Our case does not rest on government subsidies (although we also give a environmentally friendly and increasingly viable source of power can provide power supply option. scenario involving them). Instead, we look at parity between the levelized cost Delhi with an attractive long-term power supply option, reducing the city’s of solar photovoltaic (PV) systems and consumer grid prices. Assuming a reliance on power imported from other states (more than 70% of Delhi’s power fairly conservative 5% p.a. reduction in the cost of solar power and a 6% p.a. comes from outside the state) and fossil fuels (more than 50% of Delhi’s power increase in grid prices, parity will be reached in all the tariff segments by 2018. comes from coal generation). Solar will be a smart investment choice for consumers as well as for third party investors. What is needed is, above all, a level playing field. For this, grid- We believe that going solar is not so much a choice but an inevitability for 2 GW requires only connectivity of rooftop solar will be the key. Delhi. It will be driven by the increasingly favorable economics of solar and 1.6% of the city’s land. by customer demand. Entrepreneurs will develop business models to serve We arrived at our 2 GW target for Delhi by combining three perspectives. We Delhi’s power consumers as long as there are no regulatory hurdles placed first looked at the available rooftop space to reach a geographic potential. in their path. The other stakeholders, the regulators, the utilities and the Then, we looked at the economic viability of solar for different tariff groups financing community should help to actively shape this change and make it and system sizes. Lastly, we assessed how much grid-connected solar power their success, too. A good starting point for the Delhi government to show the grid could handle. For the geographic potential, data was difficult to come that it wants to lead and is ready to embrace a paradigm change in the way by, so we used different mapping sources (the Delhi Master Plan 2021, Delhi electricity is generated, distributed and consumed, is to state higher, but We believe that Development Authority data) and geo modeling of samples (individual buildings still perfectly achievable solar RPO targets. We believe that they can be at and colonies based on Google Earth, Google Maps and the Eicher map of going solar is not least 0.50% in 2014 and 9.75% in 2020 without requiring significant extra Delhi) and extrapolated from that based on assumptions. Our methodology is so much a choice but investments from the government, the utilities or the power consumers5. explained in the appendix. For the viability perspective, we used our in-house an inevitability financial models for different sized solar power plants and data from actual This report is meant to provide arguments, analysis and data to the Delhi for Delhi. solar projects in India. Assuming a fairly government, its distribution companies (DISCOMS) and the people of Delhi conservative 5% p.a. to show why solar makes sense for the city. We also provide a roadmap for The grid perspective is complex and there is hardly any reliable information reduction in the cost of achieving 2 GW of installed rooftop solar in the city by the year 2020. I hope you available. As per our knowledge, there has been no detailed study on the ability will find this report interesting. solar power and a 6% p.a. of the Delhi grid to handle distributed, intermittent power sources. Even in increase in grid prices, countries such as Germany or the USA, where studies have been conducted, parity will be reached in there is no clear indication of what a grid of specific characteristics can handle. all the tariff segments A frequently read estimate is that 15%-20% of distributed, intermittent power by 2018. in the grid is not harmful (it might even stabilize the grid). In 2013, Germany has seen days when renewable power made up more than 50% of overall power and the grid has taken it well. Dr. Tobias Engelmeier Managing Director In our analysis, we have not considered storing of solar power. Given currently [email protected] available technology, stored solar power would be economically attractive

------4 Delhi’s current economic growth rate is over 11% and is expected to remain so for 1 Installed solar capacity until December 2012. Source: Agentur für Erneuerbare the next five years. This is partially driven by its total population growth of 21% in Energien accessed at http://bit.ly/14ivVhJ on May 22nd 2013. the past decade (currently Delhi has over 16 million inhabitants); data as per Census 2 Installed solar capacity until March 31st 2013. Source: San Francisco Solar Map of India 2011 accessed at http://sfenergymap.org on May 22nd 2013. 5 As per our Roadmap for parity driven rooftop PV in Delhi - Delhi’s consumption is 3 Installed solar capacity until December 2012. Source: New York Solar Map accessed at expected to be 26,669 MUs in financial year 2014-15 and 31,057 MU in 2020-21, while http://nycsolarmap.com accessed on May 22nd 2013. solar generation potential is 149.9 MU in 2014 and 3,027.7 MU in 2020.

12 13 2 Introduction 2.1 Solar in India

Solar power plants in India in 2013 are still almost exclusively ground-mounted PV plants. Such plants are typically larger than 5 MW in size and supply electricity to the grid. Central government policies like the Jawaharlal Nehru National Solar Mission (NSM) and the Gujarat Solar Policy have provided preferential Feed-in-Tariffs (FiT) to incentivize the installation of over 1.5 GW – almost exclusively photovoltaic (PV) –in the past two years6. The NSM alone The Central Electricity wants to add another 3.6 GW by 2017. Additionally, an ever-increasing number Regulatory Commission of Indian states offer solar policies or directly allocate projects. These include (CERC) suggests that, Karnataka, Rajasthan, Tamil Nadu, Andhra Pradesh, Punjab, Madhya Pradesh, 0.25% of solar should Chhattisgarh, Uttar Pradesh and Orissa. We expect these policies to facilitate an additional installation of over 6 GW in the next four years7. be achieved by 2012-13, which should increase to In order to accelerate the growth of solar, supply-side incentives such as 3% by 2021-22. preferential FiTs are complemented by demand-side measures: Solar Renewable Purchase Obligations (RPO) define a minimum amount of solar power (MWh) that obligated entities – distribution licensees, open access and captive consumers (1 MW and above) – have to buy as a percentage of their total power sold/consumed. The Central Electricity Regulatory Commission (CERC) suggests that, 0.25% of solar should be achieved by 2012-13, which should increase to 3% by 2021-22. However, the actual targets are set by states and vary across the country. Since solar has a lower Capacity Utilization Factor (CUF) than the average of conventional sources of power, the required RPO by 2012-13 in installed capacity terms (MW) exceeds 0.25% and is nearer to 1% of the total installed capacity in India. Thus, if as of March 2013 India has 215 GW8 of installed capacity for total power generation, it needs around 2 GW of Today, the financially installed solar power to meet its cumulative solar RPOs. In 2022, if India will viable price (excluding have more than 300 GW of installed capacity and the solar RPO will be around incentives such as capital 3%, India will need nearer to 30 GW of solar power. Obligated entities can meet subsidies or tax benefits) their obligation by purchasing the required quantity of solar power directly of solar power in India from producers or producing it themselves. Alternatively, they can buy solar Renewable Energy Certificates (REC) to fulfill their RPO. is between M6.5-8.5 (€0.1-0.13, $0.13-0.17)/ Today, the financially viable price (excluding incentives such as capital kWh, depending on the subsidies or tax benefits) of solar power in India is between M6.5-8.5 (€0.1- size of the plant, cost of 9 0.13, $0.13-0.17)/kWh , depending on the size of the plant, cost of debt, cost debt, cost of land and the of land and the irradiation at the plant location10. At this price, solar power is still more expensive than the average pooled purchase cost of power, i.e. the irradiation at the average price paid by DISCOMS to purchase electricity from power generators. plant location. Solar RPO requirements and the general lack of power in many parts of the country are currently the main drivers for individual states to initiate their own solar policies and provide incentives.

------6 Source: BRIDGE TO INDIA, as of May 3rd 2013 7 Based on the BRIDGE TO INDIA market model 8 Source: Central Electricity Authority 9 Based on the BRIDGE TO INDIA financial model, excluding the seven states of the Northeast of India and Jammu and Kashmir, Himachal Pradesh and Sikkim and the cities of Delhi and Mumbai. 10 Solar for large, grid-connected plants is being offered in some states for as little as M6.2/kWh, a price currently unviable without additional tax benefits, without compromising on plant quality. Photograph by Anand Pathanjali/Greenpeace India * Rates of €1 = M65 and $1 = M50 have been used throughout this report.

14 15 In some states, however, solar has already reached parity with commercial and Apart from large, grid-connected plants, the government also supports industrial tariffs11. In addition, in many states, electricity supply is erratic and smaller, rooftop-based installations of less than 1 MW. Such support includes unreliable with widespread power cuts. To grapple with this challenge, many the Ministry of New and Renewable Energy’s (MNRE) subsidy scheme (100 consumers rely on expensive back-up power. As a result, their levelized cost MW a year), the “Solar Cities” program by the MNRE (no capacity limits), Tamil of energy (LCOE) rises and can be higher than the price for solar. There are, Nadu’s solar policy with a FiT for rooftop plants (350 MW), Kerala’s rooftop therefore, already markets where solar energy is viable without support from power program that offers a subsidy (10 MW) and the Gujarat solar policy’s the government. rooftop program that offers a FiT (30 MW). Further, rooftop PV is well suited The National Capital to serve commercial consumers and those with a high LCOE. We expect the Territory of Delhi has a Figure 1: State-wise commercial tariff vs. LCOE of solar power (100 kWp system without battery) market for such smaller PV projects without government support to exceed solar RPO set by the Delhi and diesel generation (100 kVA system) 1 GW by 201612. Electricity Regulatory 2.2 Power in Delhi and why Commission (DERC). For solar makes sense the financial year 2013- 14, this stands at 0.20% The National Capital Territory of Delhi (NCT; hereon referred to only as Delhi)13 and increases to 0.35% has a solar RPO set by the Delhi Electricity Regulatory Commission (DERC). by 2016-17. For the financial year 2013-14, this stands at 0.20% and increases to 0.35% by 2016-1714.

The city has four power distribution utilities (DISCOMS): Tata Power Delhi Distribution Limited (TPDDL), BSES Yamuna Power Limited (BYPL), BSES Rajdhani Power Limited (BRPL) and the New Delhi Municipal Council (NDMC). Together, these companies provide 100% of the power consumption needs of Delhi. Given their cumulative expected power sales of 25m MWh15 in 2013-14, they have a current requirement to purchase 50,000 MWh of solar energy, for which they require 35.5 MW of installed solar capacity16. However, the installed capacity of solar PV in Delhi as of March 9th 2013 was only 2.5 MW17. This currently leaves the Delhi utilities with the option of either purchasing solar The installed capacity power from outside the state, purchasing RECs or – when enforced – paying a of solar PV in Delhi as of penalty for non-compliance. March 9th 2013 was only 2.5 MW.

------11 Tariff groups comprise of ‘Domestic Consumers’, ‘Non-Domestic Consumers’ (all non-domestic establishments such as hotels, educational institutions, hospitals and banks), ‘Industrial Consumers’ and ‘Agriculture Consumers’. There are different tariffs for different purposes as well, such as public lighting, railway traction and advertisements and hoardings. 12 Based on the BRIDGE TO INDIA market model 13 Delhi consists of a larger metropolitan area consisting of the city of Delhi and Source: State tariff orders, market interviews, BRIDGE TO INDIA financial model and analysis adjoining satellite areas of Baghpat, Gurgaon, Sonepat, Faridabad, Ghaziabad, Noida, Greater Noida and other nearby towns. Delhi, along with the satellite cities is called the National Capital Region (NCR). The city of Delhi is known as the National Capital Territory (NCT). This report focuses on the National Capital Territory of Delhi, which is referred to as “Delhi” through the report. 14 According to the Delhi Electricity Regulatory Commission (Renewable Purchase Obligation and Renewable Energy Certificate Framework Implementation) Regulations, 2011 accessed at http://www.derc.gov.in/Main%20Page%20Matter/draft_ RPO_5_8_2011/DERC_Final_RPO_Regulations.pdf on March 13th, 2013 ------15 According to the Average Revenue Reports of 2011-12 filed by the DISCOMS with 11Tariff groups comprise of ‘Domestic Consumers’, ‘Non-Domestic Consumers’ the DERC and BRIDGE TO INDIA’s projections. (all non-domestic establishments such as hotels, educational institutions, hospitals 16 and banks), ‘Industrial Consumers’ and ‘Agriculture Consumers’. There are different At CUF of 16% tariffs for different purposes as well, such as public lighting, railway traction and 17 Source: MNRE State-wise solar installed capacity break- up - http://bit.ly/18ZeW7x advertisements and hoardings. accessed on May 14th 2013

16 17 Table 1: Delhi’s solar RPOs and required solar capacity Figure 2: India’s structural power deficit (per FY )18 2013-14 2014-15 2015-16 2016-17 Power consumption 25m 26m 28m 30m (MWh)19 Solar RPO20 0.20% 0.25% 0.30% 0.35% Required solar 35 46 60 75 capacity (MW)21

Other states in India have set higher solar RPO targets than Delhi. Haryana, Gujarat and Uttar Pradesh, for example, each aim for 1% solar power in the current financial year (2013-14). A supposed lack of available space for solar Power bought The total area of PV in a highly urbanized and congested city like Delhi is often considered to be a key barrier to a more ambitious solar RPO. However, this holds true only if we on the spot market Delhi (only the NCT) consider the installation of large, ground mounted power plants. As this report 2 can cost as much as is 1,483 km . shows, the potential for rooftop-based solar PV systems is significant in Delhi. 8 (€0.12, $0.16)/kWh. Source: CEA, BRIDGE TO INDIA projections M The total area of Delhi (only the NCT) is 1,483 km2 22. If such an area were unused and exclusively available for solar installations, it could support 123 Power bought from generators in other states has to be wheeled through GW of installed capacity which, at peak power production, would be more these states’ grids, incurring extra charges and making Delhi’s power supply than 20 times Delhi’s expected peak power demand of 6 GW23 for 2013 and dependent on their grid infrastructure and management. Power bought on the more than half of India’s installed capacity in 2013 of 215 GW24. 50% or 700 spot market can cost as much as M8 (€0.12, $0.16)/kWh. km2 of Delhi’s total area is built-up25 and theoretically available for rooftop PV systems. Based on our analysis in this report, we have estimated that of the Prices across India are going to continue to rise, driven by rising fuel costs 700 km2 of built-up space in Delhi, 31 km2 or 4.42% of Delhi’s rooftop space is (coal and diesel), a structural underinvestment into new generation capacity actually available for PV systems. This much space gives Delhi a geographic and grid infrastructure and increasing pressure on state electricity boards to Of the 700 km2 of built- solar potential of 2,557 MW. balance their books. States reeling under power deficits will increase upward up space in Delhi, 31 pressure on spot market prices to meet their own demand for short-term km2 or 4.42% of Delhi’s The power plants existing within the confines of Delhi in 2013 have a capacity peak power. of 1,345 MW (55% coal and 45% gas) . This leaves Delhi with a shortfall of 4,297 rooftop space is actually MW from its highest peak demand so far of 5,642 MW witnessed in 2012. About Figure 3: India’s rising power costs (2000-2018) By utilizing the large available for PV systems. 58% of the shortfall is met by purchasing power through long-term power potential for solar energy, This much space gives purchase agreements (PPA) with power generators in other states. A further Delhi could drastically Delhi a geographic solar 13% of the shortfall is met by purchasing power on a short-term basis from reduce its dependency on potential of 2,557 MW. other states or on the spot market. At the moment, India faces a high and rising power deficit, expected to be 13% by 2017. States that Delhi relies on for external power sources its additional power need to bridge their deficits and might be less willing in as well as its vulnerability the future to meet Delhi’s continued and growing need for power. to the kind of massive grid failure that took place in July 2012. ------18 The financial year (FY) in India runs from st1 April to 31st March. 19 According to Average Revenue Reports of 2011-12 filed by the DISCOMS and BRIDGE TO INDIA’s projections 20 Yearly targets for 2014-16 have been assumed based on the target of 0.35% by 2017 set by the DERC 21 MWh values for solar have been converted to MW by considering the power generation Source: State tariff orders, BRIDGE TO INDIA projections capacity of a 1 MW solar plant that produces power at a Capacity Utilization Factor (CUF) of 16%. CUF for solar is the ratio of actual energy generated by a plant over the year to the equivalent energy output at its rated capacity over the yearly period. By utilizing the large potential for solar energy, Delhi could drastically reduce 22 Source: Census of India 2011 provision data, accessed at www.census2011.co.in/ its dependency on external power sources as well as its vulnerability to the density.php on March 13th 2013 28 23 Source: Delhi Transmission Company kind of massive grid failure that took place in July 2012 . Further, Delhi now 24 Source: Central Electricity Authority enjoys near constant availability of power. This comes at the cost of rising 25 Source: Master Plan for Delhi 2021 electricity prices. They have increased by an average of 20% across consumer 26 Refer to the chapter on “Why rooftop solar makes sense for DISCOMS” for details on power generation sources 27 As of October 2012, Source: Delhi TRANSCO Annual Report 2011-12, more recent ------data unavailable 28 In July 2012, the country’s northern and eastern grids failed due to excess demand.

18 19 groups in the year 2012 alone29. Price hikes will likely continue as the demand in the city increases. Solar installations on the other hand, last 20 years or 3 Solar resource availability more, with the cost of power remaining unchanged for the entire lifetime of a for Delhi plant. By switching to solar, the city can hedge against future increases in grid electricity prices. Solar irradiation is the amount of radiant solar energy available per unit area and is usually expressed in terms of kilowatt-hours per square meter per day Investing into solar requires up front funds as the share of the capital (kWh/m2/day). As sunlight streams through the atmosphere, only some of it expenditure significantly outweighs the share of the operating expenditure reaches the ground, with the rest being reflected, absorbed and scattered. The for each kW. Delhi has funds. The city’s annual per capita income in 2012 was amount that actually reaches the ground depends on a number of factors such By switching to solar, M176,000 (€2,708, $3,200) – the third highest among the states and union as latitude, season, time of day, air quality and other atmospheric conditions The average 30 the city can hedge against territories and three times the national average . (e.g. clouds, aerosol particles, etc.). This can be used to generate power by way solar irradiation of of technologies such as solar PV systems. future increases in grid Solar rooftop installations are already beginning to take off across India. While Delhi according to the electricity prices. the installed capacity in 2013 is still only a little over 100 MW31, there has Table 2: Annual irradiation data for Delhi RET Screen database is been a surge in new projects and market participants. The south Indian states around 5 kWh/m2/day. especially are ready for fast growth in rooftop PV. Source RETScreen32 MNRE33 Month Solar Radiation Solar Radiation kWh/m2/d kWh/m2/d Jan 3.8 3.4 Feb 4.6 4.7 Mar 5.8 6.0 Apr 6.3 6.8 May 6.4 7.0 Jun 6.0 6.3 Jul 5.2 5.4 Aug 4.8 5.3 Sep 5.0 5.3 Oct 4.8 4.9 Nov 4.2 3.9 Dec 3.5 3.3 Annual 5 5.2 Delhi receives The average solar irradiation of Delhi according to the RET Screen database higher irradiation in is around 5 kWh/m2/day34. Such a level of solar irradiation is good, but not as relation to many other high as might be expected due to the high levels of dust in the city. The highest cities of the world that irradiation in the world is 7-8 kWh/m2/day in, for example, North Africa. The solar irradiation in Delhi is lower than that of western and southern states make a dedicated effort of India, like Rajasthan, Gujarat, Tamil Nadu, Maharashtra, Andhra Pradesh to go solar. and Karnataka, which have average irradiation of 6-7 kWh/m2/day and are considered to be the most solar suitable regions in the country. But the city of Delhi receives higher irradiation in relation to many other cities of the world that make a dedicated effort to go solar, including Berlin (98 MW installed35), Munich (27 MW36), New York (14 MW37), or Beijing (close to 15 MW installed by the end of 2013)38.

------29 Source: State tariff orders, BRIDGE TO INDIA analysis 35 Installed solar capacity until December 2012. Source: Agentur für Erneuerbare 30 Economic Survey of Delhi 2012013 accessed at http://bit.ly/19KHeBw Energien accessed at http://bit.ly/14ivVhJ on May 22nd 2013. 31 Source: MNRE, assuming off-grid projects under the subsidy mechanism are rooftop based 36 Installed until April 2012. Source: SolarBundesliga accessed at http://bit.ly/11t8Ur9 32 Source: NASA Surface meteorology and Solar Energy accessed at https://eosweb.larc.nasa.gov/ on May 22nd 2013 nd sse/RETScreen/ on May 22 2013. 37 Installed solar capacity until March 31st 2013. Source: San Francisco Solar Map 33 Source: MNRE India Solar Resource Maps accessed at http://mnre.gov.in/sec/solar-assmnt.htm accessed at http://sfenergymap.org on May 22nd 2013. nd on May 22 2013. 38 Source: www.china.org.cn/bjzt/2013-05/27/content_28923106.htm accessed on 34 India Solar Resource Map 2010 NREL May 22nd 2013

20 21 Figure 4: Irradiation maps of Delhi, Beijing and Munich 4 Delhi’s Geographic Potential for Solar Rooftop Installations

Delhi has the potential to build around 2.6 GW of solar PV on its rooftops. We have only assessed rooftop space for solar PV panels. Other technologies, such as building-integrated PV (BIPV) could additionally make walls of buildings available for solar power generation, but are still too far from commercial viability to be considered here.

Figure 5: Map of Delhi (NCT)

Source: BRIDGE TO INDIA

22 23 Delhi borders the Indian states of Haryana on the north, west and south and Figure 7: Potential of rooftop solar power generation in Delhi for different Uttar Pradesh to the east. The total land area of the National Capital Territory land area types (MW) The solar-suitable is around 1,484 km2 with around 700 km2 of built area.39 Delhi’s built area/ rooftop space is the raw rooftop space that is potentially available for solar power generation is unobstructed and around 119 km2. Raw rooftop space includes only built structures that can 40 shadow-free rooftop area accommodate the size and weight of solar installations. Out of this, the estimated, actually available solar-suitable rooftop space is around 31 km2, that receives optimal which could fit 2,557 MW of solar power41. The solar-suitable rooftop space is sunlight for solar power the unobstructed and shadow-free rooftop area that receives optimal sunlight generation. for solar power generation42.

Figure 6: Deriving Delhi’s rooftop space for solar

Source: DDA’s Master Plan 2021, Delhi Zonal Plans and BRIDGE TO INDIA analysis

Source: BRIDGE TO INDIA

Table 3: Solar suitable rooftop area available in different land area types Delhi’s residential buildings represent 49% of the solar potential. They Total Solar % of solar are followed by industrial buildings with 15% of the potential. Government qualified suitable suitable Solar buildings, commercial buildings and public and semi-public facilities have Land area rooftop area rooftop area rooftop potential Delhi’s residential 13%, 10% and 13% of the total potential, respectively. Transport facilities such type ( km2) (km2) space (MW)43 buildings represent 49% as airports and railway stations have a mere 0.1% of the potential - but can of the solar potential. make great pilot projects. Green stretches, water bodies, historical buildings Residential 74.5 14.9 49% 1,242 and public utilities have been excluded from the analysis. Commercial 10.9 3 10% 251 Industrial 11.3 4.5 15% 377 Government 10.2 4.1 13% 339 Public and 11.9 4.2 13% 346 Semi Public Facilities

------Transport 0.2 0.03 0.10% 2.6 39 As per the DDA Master Plan 2021 TOTAL 118.9 30.6 2,557 40 Raw rooftop space is computed after excluding poorly constructed or unfit building structures, roads, green patches and irrelevant sub categories such as poultry farms or religious buildings. Source: BRIDGE TO INDIA (Refer to the appendix chapter “Methodology for calculating the solar rooftop potential in Delhi” for details on the calculation.) 41 Based on Bridge to India’s estimate that 12 m2 of rooftop space is required for a 1kW solar power installation 42 Solar suitable roof top space excludes parts of the rooftop that are obstructed by objects and rooftop structures like water tanks, storage rooms, air-conditioning units, water heating units, etc. as well as parts of the roof that have a shadow. This space estimation varies for different land ------area types. (Refer to the appendix chapter “Methodology for calculating the solar rooftop potential 43 Based on Bridge to India’s estimate that 12 m2 of rooftop space is required a 1kW solar power in Delhi” for details on the calculation.) installation. 1 MW (1,000 kW) of solar requires 0.012 km2 (12,000 m2) of space.

24 25 4.1 Residential buildings belongs to type-1 and half to type-2 power consumers56. This gives each a potential of 188.5 MW. The rooftop characteristics of the two consumer types Such buildings have a solar potential of 1,242 MW, the largest among all other are assumed to be comparable. Their total, combined rooftop area is 11.3 km2 buildings types. Residential buildings include those under the jurisdiction of (after discounting for old buildings that might not be able to support solar). Out the Municipal Corporation of Delhi (MCD), the New Delhi Municipal Council of this, the solar suitable area after discounting for rooftop obstructions and (NDMC) and the Delhi Cantonment Board (Delhi Cantt.). The colonies under shadowing is assumed to be 40%. Accordingly, the total solar suitable rooftop the MCD are further categorized from A to H, as per the circle rate44 applicable space is 4.5 km2 57. to each colony. Colonies under the categories E, F, G and H are excluded from our analysis, as they were unsuitable for solar power generation45. The total 4.4 government buildings built area available for rooftop solar installations is around 74 km2 46. The solar suitable rooftop space is around 15 km2 47. To arrive at this number, the solar These buildings have a solar potential 339 MW. The total qualified built-up area of government buildings is around 10 km2. The solar suitable rooftop space The total solar potential suitable rooftop space was assumed to be around 20% of the total rooftop area48. This percentage is based on samples and discounts for the space offered by government buildings is assumed to be around 40%, which is larger The total qualified of commercial buildings occupied and shadows created by water tanks, ventilating shafts, dry clothes than many other land area types, considering their rooftop space is commonly built-up area of is around 251 MW. lines and storage units that occupy most of the area on a typical residential free from elevator shafts and storage units58. Accordingly, the solar suitable government buildings is 49 2 rooftop . rooftop space is 4 km . around 10 km2. 4.2 commercial buildings 4.5 Public and semi-public facilities

The total solar potential of commercial buildings is around 251 MW. Public and semi-public facilities can be government owned or privately owned. Commercial consumers can be categorized into two types, those with an Accordingly, they pay different prices for electricity and the viability of solar electricity consumption of 10-200 kW (“type-1”) and those with a consumption differs59. Based on market interviews and sampling, we assume that 70% of the of greater than 200 kW (“type-2”). The two types of consumers pay different facilities are government owned and 30% privately owned60. Such facilities have tariffs and thus the viability of solar differs50. Their total built-up area is a total potential of generating 346 MW of solar power: 242 MW on government 11 km2 51. Rooftops of commercial buildings tend to be occupied with cooling buildings and 104 MW on private buildings. The total qualified rooftop area is towers, ventilating shafts, water heating systems and tanks. The solar suitable around 12 km2. Out of this area, the solar suitable rooftop space is around 4 rooftop space for type-1 commercial buildings is assumed to be 30% and that km2 61. The solar suitable rooftop space on public facilities is assumed to be of type-2 is 20% of their respective rooftop areas. The total solar suitable 35%, taking into consideration that rooftops of most public facilities are largely rooftop space for commercial buildings is 3 km2 52. occupied by water heating systems (mandatory as per government regulations) and elevator and ventilating shafts. In addition to that, the structure of most A few large-sized, commercial buildings have already installed rooftop solar buildings in this land area type is such (elevated at some levels) that significant power installations in order to meet part of their power needs, as solar area is shadowed. In the following analysis, the government-owned facilities power is economically viable for such users. For example, DLF promenade, are included in the overall “government” potential, while the privately-owned a shopping mall in Vasant Kunj with a total raw rooftop space of 9,000 m2 53 facilities are included in the “commercial (type-2)” potential. and solar suitable roof space of around 1,800 m2, has installed a 50 kW solar rooftop installation54. 4.6 tRAnsport facilities Transport facilities Industrial buildings have a conservatively 4.3 Industrial buildings This land area type has a conservatively assessed solar potential of only assessed solar potential have a solar potential of 2.6 MW. The rooftop space available is around 0.2 km2. The solar suitable 377 MW. Industrial buildings have a solar potential of 377 MW. Like commercial rooftop space is around 0.03 km2 62. The solar suitable roof area is assumed of only 2.6 MW. customers, we also categorize industrial customers into two types: those to be 20% of the total raw rooftop area as most transport facilities have roof with 10-100 kW of load requirement (“type-1”) and those with 100-200 kW of sheeting and the little concrete space left is occupied by shafts and monitoring load requirement (“type-2”). The two types also pay different tariffs, which units64. Since this sector has some landmark buildings such as the Delhi impacts the financial suitability of solar power55. Based on market interviews Airport (where 2 MW of solar is to be built on the surface around the runways), and due to a lack of specific data, we assume that half of the rooftop space we have included it in the geographic analysis. However, because the solar ------potential is negligible compared to other segments, we have excluded it from 44 Circle Rate is the minimum rate per sq. meter area on which stamp duty has to be paid by the further viabilty analysis and computation of the overall potential. the purchaser. This rate is decided by the government authorities for valuation of land in a particular area. ------45 Refer to Table A2 “Assessment of colonies under the jurisdiction of the MCD” for details 56 Refer to the appendix chapter “Computation of solar suitable rooftop space – 46 Refer to the appendix chapter “Computation of the solar suitable rooftop space - Industrial buildings” for details 57 Residential buildings” for details Refer to the appendix chapter “Computation of the solar suitable rooftop space - Industrial buildings” for details 47 Ibid, 48 Ibid. 49 Ibid. 58 Refer to the appendix chapter “Computation of the solar suitable rooftop space - 50 Refer to Chapter 5 on “The viability of rooftop solar PV” for details Government buildings” for details 51 Refer to the appendix chapter “Computation of the solar suitable rooftop space - 59 Refer to Chapter 6 on “The viability of rooftop solar PV” for details Commercial buildings” for details 60 Refer to the appendix chapter “Computation of the solar suitable rooftop space – 52 Ibid. 53 As per measurements taken from Google Earth Public and semi –public facilities” for details 54 Refer to the Appendix for details on the approach and assumptions of the analysis 61 Ibid. 62 See appendix for details 55 For details, refer to Chapter 5 on “The viability of rooftop solar PV” 63 Refer to the Appendix for details on the approach and assumptions of the analysis

26 27 5 Delhi’s existing solar 6 Integration of solar PV policies and incentives with the grid

There exists no Delhi solar policy in 2013. Delhi had formulated a draft rooftop Integration of a rooftop solar PV system with the grid is not a technical solar policy in 2011 to promote rooftop, small-scale, decentralized solar power requirement as it could also provide a purely captive solution, where the entire generation to meet its solar RPO. The policy envisioned that homeowners power generated could either be consumed immediately or stored for later could either lease their roofs to a developer or install a system themselves. use at the site of generation. However, grid connection can be a key driver for The government was planning to offer a FiT of M17 (€0.26, $0.34)/kWh for the rooftop PV as it allows systems to avoid storage (which can increase system power generated by the PV plant. We assume that at the time the cost of a 5 costs by 25% or more) and offers more off-take options (i.e. the grid through, Power grids were kW system without storage was around M900,000 (€13,846, $18,000).64 for example, net-metering or FiTs). originally designed for The Delhi government However, the Delhi rooftop solar policy was not implemented. In 2011, the the one-way supply of Power grids were originally designed for the one-way supply of power from a currently has no specific economics were not favorable. The cost of diesel power was only M12 (€0.18, power from a central 65 central source of power generation to the consumers. The security, control, policy for supporting $0.24)/kWh , which was significantly lower than the proposed solar FiT. As a result, the government was concerned that cheaper diesel power would protection, power flows and earthing of the network were based on this source of power 71 rooftop solar power. be fed into the grid, disguised as solar power, in order to artificially inflate centralized generation model. Distributed generation technologies , such as generation to the the revenue of the PV system. Today, however, such concerns are no longer solar PV systems, when connected to the larger distribution network of the consumers. valid. The cost of diesel power as of May 2013 has risen to M14 (€0.21, $0.28)/ grid create a two way supply of power, from the grid to the consumer and from kWh66, an increase of 17% since October 2011. The capital cost today for a 5 the consumer to the grid. This requires a re-thinking of the way the grid is kW system has fallen by 39% to M550,000 (€9,231, $12,000). At this cost, a FiT regulated and managed. of M11.6 (€0.17, $0.23) /kWh is viable67, which is lower than the cost of diesel power today. Figure 8: Centralized vs. mixed (centralized and distributed) power generation The Delhi government currently has no specific policy for supporting rooftop solar power. As part of the first phase of the NSM, central government support has been available in the form of the 30% MNRE subsidy on the PV system cost or a low cost loan at 5% interest, repayable within five years. Such support was applicable up to March 31st 2013, and has been offered in the draft policy for the second phase of the NSM as well. Pending finalization of the draft, there is no certainty yet on whether the support will continue in the second phase. The central government also provides direct tax benefits. One such benefit is A rooftop solar policy accelerated depreciation of 80%68, attractive to profit making business entities, by Delhi can set a paying tax in India that own a PV system. Another is a 10-year income tax exemption for PV power generators and indirect tax benefits such as excise benchmark by creating duty and custom duty exemptions or concessions on solar energy devices and a policy that addresses equipment. the need for grid- connectivity and achieves The central government support for rooftop solar has been successful but limited: support was initially pledged only for 100 MW (124 MW has actually significant capacity been installed in India)69. It is available only for off-grid systems. There is no addition (up to 2 GW), at policy framework to allow rooftop systems to connect to the grid, something the limited cost. that can impact viability significantly. A rooftop solar policy by Delhi can set a benchmark by creating a policy that addresses the need for grid-connectivity and achieves significant capacity addition (up to 2 GW), at the limited cost70.

------64 This capital cost is an assumption based on our historical understanding of the market in 2011. There is no publically available information on the capital cost considered by the government for its FiT in 2011. Source: BRIDGE TO INDIA 65 For a 15 kVA system that consumes 3.5 liters of diesel/hour. Considering diesel prices to be M40.91/liter as on October 2011 66 For a 15 kVA system that consumes 3.5 liters of diesel/hour. Considering diesel prices to be ------M48.63/liter as on April 2013 71 Distributed generation technologies are small power generating units that are placed at or 67 Refer to the section on “Viability of solar rooftops for residential consumers” in Chapter 5 of near the point of energy consumption, unlike traditional “centralized” systems, where electricity this report is generated at a remotely located, large-scale power plant and then transmitted down power 68 On the book value of assets in the first year and the written-down value thereafter, under section lines to the consumers. Distributed energy encompasses a wide range of technologies including 32 Rule 5 of the Income Tax Act wind turbines, solar PV, fuel cells, micro turbines, reciprocating engines, load reduction 69 Source: MNRE technologies, and battery storage systems. Source: www.nrel.gov/learning/eds_distributed_ 70 Refer to the chapter “Delhi solar roadmap: reaching 2 GW by 2020” in this report for more details energy.html accessed on May 22nd 2013

28 29 capacity in Delhi that can be connected to the grid. However, this “grid limit” is recognized as being conservative. Simple, low-cost measures like introducing specific standards for PV inverters, guidelines for grid connectivity and better load forecasting and planning can make a difference in the amount of PV penetration possible in Delhi. The Central Electricity Authority (CEA) has already drafted a report on the integration of solar PV systems in to the grid that addresses such measures74. The authority is developing more detailed recommendations in its final report, which is currently being assessed by the forum of regulators and could be released by late 2013. By adopting the We estimate that the recommendations of the report and by better forecasting demand, we estimate possible PV penetration that the possible PV penetration limit in Delhi would be at least 20%75. Beyond a certain limit limit in Delhi would be at of PV penetration, In the future, the grid’s capacity for PV can be increased drastically if the grid least 20%. is upgraded with smart grid management/monitoring solutions (including traditional grids can advanced forecasting, scheduling and load management) and centralized or be destabilized. decentralized storage76. Countries like Germany, with an already advanced grid infrastructure, have reached 40% of decentralized PV in the power mix by adding intelligent transformers and storage devices that regulate the quality of power and stabilize the grid77. Such solutions increase the hosting capacity of the grid, but require significant investments.

Source: BRIDGE TO INDIA

Traditional grids are not designed for a two way supply of power. Beyond a certain limit of PV penetration, traditional grids can be destabilized due to e.g. the intermittent quality of solar power, reverse power flow (when a PV system at the consumers end generates more power than the consumer can use) and voltage imbalances72.

Decentralized PV systems can alter the load profile of the network as the power from such systems can be consumed on site as well as injected in to Decentralized PV systems the grid. High on-site PV consumption by consumers could reduce the load on can alter the load profile the network during the day, while increasing it at night (because of an absence of the network as the of PV generation). Further, the load on the grid can also fluctuate as PV power from such systems generation is highly weather dependent and can drastically drop in the case of can be consumed on site a cloud cover. High levels of grid-connected decentralized PV penetration can present a significant load management challenge for the DISCOMS. as well as injected in to the grid. As per estimates, a traditional grid can accommodate up to 15% of the peak load from distributed solar PV without being destabilized, while managing load fluctuations easily73. This places a “grid limit” on the rooftop solar PV ------74 Source: CEA “Integration of solar systems with thermal/hydro power stations and connectivity of solar rooftop systems with the grid” accessed at http://www.cea.nic.in/more_upload/task_ force_solar_report.pdf on May 22nd 2013. ------75 According to the NREL study on “Interconnecting PV on New York City’s Secondary Network 72 Refer to the Appendix, chapter “Solar PV grid connectivity challenges and solutions” Distribution System”, PV penetration levels of 20-30% are considered safe for radial distribution for further details. grids, like the one in Delhi. The study was accessed on May 22nd 2013 at http://www1.eere.energy. 73 The US Federal Energy Regulatory Commission recommends that PV generation should not gov/wip/pdfs/46902.pdf . The CEA, DISCOMS in Delhi and leading energy experts have shared this exceed 15% of the annual peak load. This is based on a rationale that voltage deviations due to view with us while conceding the need for a more detailed study on the matter, specific to Delhi. the intermittency of solar and the problems specific to grid integration of distributed PV have 76 For more details about challenges of PV penetration and suggested solutions see the report a negligible impact on the grid, if the combined PV power on a line section is always less than the “Connecting the Sun“, November 2012 by the European PV Industry Association accessed at http:// minimum load. The minimum load usually does not exceed 30% of the peak load. Therefore, 15% bit.ly/13bL9lH accessed on May 22nd 2013. or half of the minimum load is taken as a conservative number acceptable to ensure that there is 77 PV systems have contributed to up to 40% of peak demand in Germany (http://bit.ly/11mwWko) no risk of PV capacity having a negative impact on the grid. Source: https://solarhighpen.energy. “High Penetration PV Cases in California, US” http://1.usa.gov/185cGvY both sites accessed on gov/article/beyond_15_rule. May 22nd 2013

30 31 since 200479. Secondly, Delhi has seen a drive to making power available to 7 the viability of rooftop the city almost 24/7. As a result, DISCOMS need to match peak demand and solar PV manage demand fluctuations, which often lead to wastage of power reserves and the need to buy more high cost (peak) power. Based on the geographic analysis in chapter 1, Delhi’s electricity consumers can be divided into four categories: commercial, industrial, residential and Figure 9: Historical grid tariff price trend80 government. Together, these customer groups can accommodate 2.5 GW of solar on their rooftops. Commercial consumers include type-1 commercial There are three types of buildings and type-2 commercial buildings combined with privately owned public and semi-public facilities. Industrial (type-1 and type-2) and residential electricity tariffs in Delhi: buildings are each a consumer category in themselves. Finally, government non-domestic, industrial consumers include government buildings and public and semi-public facilities and domestic. owned by the government.

There are three types of electricity tariffs in Delhi: non-domestic, industrial and domestic. Commercial consumers pay non-domestic prices. Industrial consumers pay industrial prices. Residential and government consumers pay domestic prices.

Table 4: Consumer categories for rooftop solar PV

Building types Potential Tariff (M/kWh) (MW) Commercial Commercial type-1 205 8.5 (€0.13, $0.17) consumers Commercial type -2 149 7.15 (€0.11, (type-2 commercial $0.14) buildings and Source: Aggregate revenue requirements privately owned (ARRs) of Delhi DISCOMs public and semi- N on-domestic public combined)

Industrial Industrial type-1 189 6.6 (€0.10, $0.13) While grid power prices have risen, the cost of solar power has declined consumers Industrial type-2 189 8.35 (€0.13, sharply in the past two years. At the beginning of 2011, a ground mounted While grid power $0.17) MW-scale project in India cost around M140m (€2.15m, $2.8m) per MW. Today, prices have risen, the Industrial this has fallen by almost 50% to around M70m (€1m, $1.4m) per MW. The cost cost of solar power has reduction has been driven largely by a drop in module prices from M50 (€0.77, Residential Residential 1,242 6.4 (€0.09, $1.3) $1)/Wp in 2011 to M32.5 (€0.5, $0.65)/Wp today. Module prices have fallen declined sharply in the consumers globally due to a condition of acute oversupply. Against global demand for 30- past two years. Government Government 582 6.4 (€0.09, $1.3) 35 GW of solar power in 2012, available supply was 50-60 GW, most of it 81 facilities buildings and from China.

Domestic government owned public and semi- public

While power prices in Delhi have kept quite constant between 2005 and 2011, they have increased steeply since 2011. During the entire interval, domestic prices have increased from M4.65 (€0.07, $0.09)/kWh in 2008 to M6.4 (€0.98, $0.13)/kWh in 2013, a rise of 27%. Similarly, industrial prices have increased by 37% and non-domestic prices have increased by 33% during the same period78. There are two main drivers for these price rises. Firstly, there has been a 50% ------79 rise in the cost of international procurement of coal for Indian power producers Source: ARRs of Delhi DISCOMs 80 Tariff orders by the Delhi Electricity Regulatory Commission (DERC) accessed at http://www.derc.gov.in on May 22nd 2013 ------81 Source: Reuters article “China's bailouts darken horizon for solar panel sector” accessed at 78 Source: ARRs and tariff orders of Delhi’s DISCOMS http://reut.rs/WUn1no on May 22nd 2013

32 33 Figure 10: Falling prices of solar PV modules over the past 32 years Figure 12: Solar costs in India in 2013 for different system sizes (without storage)

Source: BRIDGE TO INDIA analysis. Source of data NREL82

Figure 11: System prices have fallen over 47% in last 10 years

Source: Market interviews, BRIDGE TO INDIA financial model and analysis.

Source: Module prices data from Renewable Energy World83, BRIDGE TO INDIA analysis

The unusually rapid fall in solar prices in the recent years and an increase in With such high investment costs, rooftop solar needs to be financially viable for Rooftop solar needs to The unusually grid prices has made solar competitive. However, we assume that now the cost consumers investing directly in to their PV plants, or for businesses investing be financially viable for rapid fall in curve will again follow its long-term trajectory. The timing for investing into into providing solar energy as a service to end consumers. Viability means that consumers investing solar prices in solar is, therefore, ideal. the investment into a solar PV plant can provide a return on investment akin to alternate investment opportunities. directly in to their PV the recent years and While increasingly competitive on a per kWh basis, solar still requires plants, or for businesses an increase in significant upfront capital investments and hence liquidity. Capital expenditure The graph and table below summarize the viability of solar for different investing into providing grid prices has made on PV plants can range from M550,000 (€8,462, $11,000) for a typical 5 kW consumer types, followed by a detailed analysis. solar energy as a service solar competitive. system for residential consumers to M140m (€2m, $2.8m) for a 2,000 kW, to end consumers. bundled installation for government consumers (cost per Wp reduces as system sizes increase).

------82 Source: NREL accessed at http://www.nrel.gov/docs/fy12osti/51847.pdf on May 22nd 2013 83 Source: Renewable Energy World accessed at http://bit.ly/WimCXG on May 22nd 2013

34 35 Figure 13: Delhi rooftop solar market size and viability Cost of solar (INR/Wp) 110 100 90 100 90 70 Cost of a typical solar 0.55 2.50 4.50 2.00 3.60 140 system (m INR)

Grid tariff in 2013 (INR/ 6.4 8.5 7.15 6.6 8 6.4 kWh)

Capacity Utilisation 16% 16% 16% 16% 16% 16% Factor (CUF) (in %) O&M expenses (INR/ 5,000 25,000 35,000 20,000 40,000 6,00,000 year/entire average system)

Equity:Debt 100% equity 30:70 30:70 30:70 30:70 30:70 Loan period (years) n.a. 12 12 12 12 12 Cost of debt (%) 13% 13% 13% 13% 13% 13%

Source: BRIDGE TO INDIA Required Equity IRR (%) 15% 15% 15% 15% 15% 15% financial model and analysis Viability Analysis 2013 Achieved equity IRR (%) 7.75 12.15 11.18 7.72 12.9 14.55 Required price for solar 11.6 9.7 8.6 9.7 8.8 6.53 per kWh in 2013 (@ 15% EIRR) Viability gap (INR per 5.2 1.2 1.45 3.1 0.8 0.13 kWh) Viable in 2013? NO NO NO NO NO YES

Viability 2013-2020 Table 5: Viability of rooftop solar PV in Delhi in 2013 (additional input factors) Commercial Industrial Assumed annual grid 5% 5% 5% 5% 5% 5% Type 1 Type 2 Type 1 Type 2 tariff escalation (%) Residential (medium) (large) (micro) (small) Government Assumed annual 6% 6% 6% 6% 6% 6% Geographic reduction in solar costs Potential (%) Total potential (MW) 1,242 205 149 189 189 582 Parity year reached 2018 2014 2015 2016 2014 2013

Input Factors 2013 Note: Straight-line depreciation and net-metering has been assumed for all consumers . Typical rooftop size 200 to 1,000 400 to 2,000 2,000 to 300 to 800 800 to 1,500 1,000 to Source: Market interviews, BRIDGE TO INDIA analysis and financial model range (sq m) 12,000 13,000 Solar suitability of 20% 30% 20% - 35% 40% 40% 35%-40% rooftop (%) Typical solar potential 3 to 17 10 to 50 33 to 350 10 to 27 27 to 50 29 to 433 per rooftop (kW) Typical system size (kW) 5 25 50 20 40 2,000 System size cost 5 to 14 15 to 34 35 to 74 15-34 35-74 500 to 2,000 bracket (kW)

36 37 Figure 14: Viable solar tariff for commercial consumers vs. non-domestic 7.1 Viability of rooftop solar PV for 88 commercial consumers grid tariff (type-1 and type-2)

Commercial consumers pay the non-domestic tariff for electricity, which can be as high as M8.5 (€0.13, $0.17)/kWh84. Based on their load requirement, The viable solar tariff commercial consumers fall into two tariff categories. Type-1 consumers (including a return (e.g. a small office complex) with a load demand between 100-200 kW pay a expectation of 15%) higher tariff of M8.5 (€0.13, $0.17)/kWh. Type-2 consumers (e.g. a mall like DLF promenade, a privately owned hospital like Apollo Hospitals) with a load for a 25 kW solar plant demand higher than 200 kW pay a lower tariff of M7.15 (€0.11, $0.14)/kWh. constructed in 2013 is M9.7 (€0.15, $0.18)/kWh, Based on the load requirements and available rooftop sizes, we considered which is only slightly an average system size of 25 kW for type-1 consumers and 50 kW for type-2 85 higher than the current consumers . The viable solar tariff (including a return expectation of 15%) for a 25 kW solar plant constructed in 2013 is M9.7 (€0.15, $0.18)/kWh86, which grid tariff of M8.5 (€0.13, is only slightly higher than the current grid tariff of M8.5 (€0.13, $0.17)/kWh. $0.17)/kWh. Type-1 commercial consumers reach parity in 2014. Solar will be viable for them as it will provide returns comparable to the returns a typical small or medium business would aim to generate in India (16-18%) and higher than the cost of debt (13%).

For a 50 kW solar PV system the viable solar tariff is M8.6 (€0.13, $0.17)/kWh87, whereas the current grid tariff is M7.15 (€0.11, $0.14)/kWh. The difference in solar tariff is due to the scale effect that lowers the per kW cost. Solar power For a 50 kW solar PV is still more expensive than grid power for type-2 commercial consumers. system the viable solar They will reach parity only in 2015. Solar will then be an attractive investment tariff is M8.6 (€0.13, proposition. Source: DERC tariff orders, BRIDGE TO INDIA analysis $0.17)/kWh, whereas the current grid tariff is M7.15 (€0.11, $0.14)/kWh. 7.2 Viability of rooftop solar PV for industrial consumers

Delhi has no registered medium to large-scale industries89 – only small and micro-scale industries.90 Based on their electricity load, they pay different electricity tariffs. Type-1 consumers (e.g. small-scale textile units and printed presses) have a load requirement between 10–100 kW. Type-2 consumers The equity internal rate (e.g. auto-parts manufacturers and chemicals manufacturers) have a load of return (IRR) of an requirement between 100-200 kW. ------investment into a solar 84 Source: Annual Revenue Requirements of Delhi’s DISCOMS PV system for 85 Refer to the chapter in the appendix on “Rooftop solar PV systems” for details For type- 1, for the purpose of our analysis, we are considering a PV system 86 As per the BRIDGE TO INDIA financial model. We are assuming that there will be net-metering, size of 20 kWp and 40 kWp for Type- 291. The viable solar tariff for a 20 kW a type-1 consumer today providing an off-take for solar power also on Sundays and holidays, when commercial 92 establishments like small offices will be shut. On working days, the consumers will completely system constructed in 2013 is M9.7(€0.15, $0.19)/kWh , which is significantly would be 7.72%, lower consume solar power on site as their load (100-200 kW) is more than the assumed standard higher than the current grid tariff of M6.6 (€0.10, $0.13)/kWh. The equity than the typical IRR installed solar capacity (25 kW). internal rate of return (IRR) of an investment into a solar PV system for 87 expectation of 16-18% for As per the BRIDGE TO INDIA financial model. We are assuming that there is a complete daily 93 off-take of the solar power, i.e. that the power is consumed also on Sundays and holidays, as this a type-1 consumer today would be 7.72% , lower than the typical IRR a business in India. is usually the case for commercial establishments like hospitals and malls. Further, the expectation of 16-18% for a business in India. Type-1 consumers reach parity consumers will completely consume solar power on site as their load (greater than 200 kW) is more than the assumed standard installed solar capacity (50 kW). only in 2016, at which point solar will become a viable investment option. 88 Assuming a 5% annual rise in electricity and 6% fall in solar system costs per year 89 Classification of industries by the government of India is based on the total investment into machinery or services. An industry, which invests less than M2.5m (€38,462, $50,000) into machinery or M1m (€15384, $20,000) into services is a “micro scale industry”. An ------91 industry, which invests between M2.5m (€38,462, $50,000) and M50m (€769,230, Refer to the chapter in the appendix on “Rooftop solar PV systems” for details $1,000,000) into machinery or between M1m (€15,384, $20,000) and M20m 92 As per the BRIDGE TO INDIA financial model. We are assuming that there will be net-metering, (€307,692, $400,000) into services is a “small scale industry”. providing an off-take for solar power also on Sundays and holidays, when industries might be Source: www.dcmsme.gov.in/faq/faq.htm accessed on May 22nd 2013 shut. Net metering will also provide an off-take in case a consumer’s load is lower than the 90 Source: Ministry of Small and Medium Enterprises accessed at assumed standard solar installation of 20 kW. http://bit.ly/1b73m78 on May 22nd 2013. 93 As per the BRIDGE TO INDIA financial model.

38 39 Figure 15: Viable solar tariff for industrial consumers vs. industrial grid A small-scale rooftop system of 5 kW (average) for a residential consumer97 tariff (type-1 and type-2) costs M550,000 (€8,461, $11,000)98 or M110 (€1.69, $2.20)/Wp. At this cost, the viable solar tariff for a residential consumer is M11.60 (€0.18, $0.23)/kWh99. This is 81% higher than the grid tariff of M6.4 (€0.09, $0.12)/kWh.

If investing in a PV plant today, a residential consumer can earn an equity The viable solar tariff for IRR of only 7.75%100. This return is not much higher than the inflation rate of a residential consumer 5% and falls short of other low-risk investment options like fixed deposits or is M11.60 (€0.18, $0.23)/ mutual funds that can give returns of 9% and above, making solar unviable for residential consumers today. kWh. This is 81% higher than the grid tariff of M6.4 Figure 16: Residential solar investment compared to similar low-risk (€0.09, $0.12)/kWh. investment options101

Source: DISCOM tariff orders, BRIDGE TO INDIA analysis

The viable solar tariff of a 40 kW solar PV system in 2013 is M8.8 (€0.14, $0.18)/kWh94, which is slightly higher than their current grid tariff of M8 (€0.12, $0.16)/kWh. Type- 2 industrial consumers reach parity in 2014, when the equity IRR for an investment into a solar power plant would be 15%95. As in the case of commercial consumers, the equity IRR for type- 2 industrial consumers is Type- 2 industrial comparable to the returns a typical small or medium business would aim to consumers reach parity generate in India (16-18%) and higher than the cost of debt (11-13%). Thus, in 2014, when the equity we believe that for these customers, PV is a sufficiently attractive investment proposition. IRR for an investment into a solar power plant 7.3 Viability of rooftop solar PV for would be 15%. residential consumers

Residential consumers in Delhi pay the domestic tariff for electricity, which is Source: Market interviews, BRIDGE TO INDIA analysis M6.4 (€0.09, $1.3)/kWh96. It is the segment with the highest rooftop potential for solar installations at 1.24 GW.

------97 Refer to the chapter in the Appendix on “Rooftop solar PV systems” for details 98 According to our market information, as of May 2013, a 1 kW PV system costs M120,000 (€1,846, $2,400) ------99 As per the BRIDGE TO INDIA financial model. We assume net metering is available, providing 94 As per the BRIDGE TO INDIA financial model. We are assuming that there will be net-metering, an off-take to the consumer in case their load is below the assumed standard solar installation providing an off-take for solar power also on Sundays and holidays, when industries might be shut. of 5 kW. On working days, the consumers will completely consume solar power on site as their load of 100- 100 As per the BRIDGE TO INDIA financial model. 200 kW is more than the assumed standard installed solar capacity of 40 kW. 101 Investment of M550,000 (€8,461, $11,000) required for a typical 5 kW residential system 95 As per the BRIDGE TO INDIA financial model. has been considered for analysis. Returns calculated after 7 years. 20% tax on fixed deposit. 96 Tariff for residential consumers who consume above 400 units per month Return on gold calculated based on past 15 year data

40 41 In addition to the IRR, the payback period, i.e. the amount of time it will take Figure 18: Viable solar tariff for residential consumers vs. At today’s prices, for the system owner to recover his cost, is also an important decision-making domestic grid tariff residential solar has a factor for a household. Delhi is a rapidly developing city where buildings might payback period of undergo major construction work every decade. The PV system needs to at least realize a payback for the owner before the owner moves or the building is 12 years. structurally changed. At today’s prices, residential solar has a payback period of 12 years.

Figure 17: Payback from investment in a 5 kW residential system (2013)

Source: ARRs of Delhi DISCOMs, BRIDGE TO INDIA analysis

down the cost per kW significantly, giving government buildings a strong viability push.

Government buildings are currently paying the domestic consumer tariff of M6.40 (€0.09, $0.12)/kWh. The viable solar tariff for a 2,000 kWp system is currently M6.53 (€0.10, $0.13)/kWh106, making solar already viable for government consumers in 2013. Source: BRIDGE TO INDIA financial model and analysis Figure 19: Viable solar tariff for government buildings vs. domestic grid tariff

For solar to be acceptable to residential consumers, a payback period of 5-7 years is preferred102. A seven-year payback for 5 kW PV system also corresponds to an equity IRR of 15%103, which would make returns from solar significantly higher than alternate, low-risk investments. Under these conditions, residential solar will reach parity in 2018. 7.4 Viability of rooftop solar PV for For solar to be acceptable government facilities

to residential consumers, Delhi has several government buildings that belong to the central government, a payback period of 5-7 state government and city administration. The major advantages of years is preferred. government buildings are that they are located next to each other, typically in the same complex, and that they have large rooftops (e.g. Krishi Bhawan in Delhi could accommodate 430 kW and Nirman Bhawan 330 kW104). The proximity between buildings facilitates the bundling of projects. Hence project 105 Source: DERC tariff order, sizes (of bundled sites) of 2,000 kWp and above become possible . This brings BRIDGE TO INDIA analysis

------102 ------Based on market interviews 106 103 As per the BRIDGE TO INDIA financial model. We are assuming that there will be net-metering, Source: BRIDGE TO INDIA financial model providing an off-take for solar power also on Sundays and holidays, when some government 104 As per BRIDGE TO INDIA analysis and mapping of rooftops through Google Earth buildings might be shut. Net metering will also provide an off-take in case consumers’ loads are 105 Refer to the appendix chapter “Rooftop solar PV systems” for details lower than the assumed bundled solar installation of 2,000 kW.

42 43 of solar through collective buying of solar PV systems. 1BOG first approaches 8 Business models for systems installers (and module manufacturers in some cases) to negotiate a rooftop solar discount for a significantly larger capacity than an individual household could. It then encourages homeowners to go solar and sign up for the discount by educating them (through the website) about the costs and benefits associated 8.1 Straight-forward sales model with a solar system installation.

The most common business model for solar deployment in India today is to 1BOG, acts as an agent that connects buyers to sellers. It charges a referral simply sell either a complete plant, usually by an Engineering, Procurement fee from the system installer. The first 1BOG initiative111, enabled around 41 and Construction company (EPC), or individual components (such as modules citizens of San Francisco to go solar with a group discount of 20%. or inverters) to an installer or end customer. The plant owner pays 100% of the PV system cost upfront. This model (sometimes referred to as the “CAPEX107 In order to organize communities to sign up for the pre-arranged discount, The main drawback of model”) is pursued by the majority of solar companies, including TATA Power 1BOG runs “campaigns”. These campaigns encourage rooftop owners to Solar, EMMVEE Photovoltaics or Moserbaer. the CAPEX model is that bring in additional participants to receive further discounts. The benefits the plant owner needs of the discounts are availed by communities as a whole, thus motivating all The model can be extended by, for example, developing specific, standardized to be able to finance the participants to spread the word. The campaign lasts for a few months, during solar kits108 for specific customer types perhaps in the range of 1-50 kWp. which the company canvasses, employs social media marketing techniques entire plant. Consumers can purchase these kits from suppliers and have them installed by and uses local advertising (such as placards or lawn/road signs) to a local installer109. capture customers.

The main drawback of the CAPEX model is that the plant owner needs to 1BOG also encourages households in states that do not yet have solar targets be able to finance the entire plant. Solar has a heavily “front loaded” cost or solar incentives to push lawmakers and local government authorities to structure, with a high initial investment and very low operating costs. A create solar policies to reduce the carbon footprint and create green jobs. customer might not have the required liquidity to finance a system upfront or An interactive map on the 1BOG website allows power consumers to enquire get the best debt terms. about discounts available in their state. If the particular state does have any policies supporting solar, the consumer is urged to contact government A key USP of this model is that it allows industrial and commercial consumers authorities to push for solar112. 1BOG now operates in over 40 states in the to own the system and claim tax depreciation benefits. With regards to smaller United States113. consumers, this business model has not achieved scale in India yet. This might be due to the cost. However, a lack of customer awareness and sales channels also plays a key role. Another concern is that the customer who buys the plant 8.2 Renewable Energy Service In a city such as Delhi, might not always have the skills to maintain it properly. Company (RESCO) Model where the household income is high, such a In a city such as Delhi, where the household income is high, such a model Under a RESCO model, a third party investor comes in to invest into a PV The RESCO model is could work with residential consumers as well. Standardization, aggressive plant on a rooftop and sells solar power to a power consumer. There is no model could work with applicable particularly marketing and information campaigns, easy availability of PV equipment, investment required from the consumer. If solar power is viable, as discussed to such commercial and residential consumers credible product guarantees and operation and maintenance plans as well as before114, the consumer can benefit from savings on the electricity bill right industrial consumers in as well. consumer finance are needed to provide a conducive market environment. The from the start. Under this model, the investor and the consumer agree on a model can be bolstered through e.g. the emergence of rooftop aggregators tariff (per kWh of solar power) and timeline of a power purchase agreement Delhi who are not willing or agents that can connect customers and installers, offering lower PV costs (typically between 15-25 years). Operations, repair and maintenance of the or able to invest e.g. to groups of customers and functioning as sales channels for installers or system could be carried out by the RESCO company rather than the consumer. M2.5m (€38,461, $50,000) equipment manufacturers. An example of this is the “One Block off the Grid” for a 100 kW system, initiative in the US. This model becomes more feasible with larger project sizes (either of individual large plants or of many small one’s bundled together) as the but are ready to commit Case Study transaction costs would prevent most investors from investing into individual to an attractive long- small plants. The model is applicable particularly to such commercial and term power purchase “One Block Off The Grid” (launched in 2008) industrial consumers in Delhi who are not willing or able to invest e.g. agreement. M2.5m (€38,461, $50,000) for a 100 kW system, but are ready to commit to an “One Block off the Grid”110 (1BOG) is a US for-profit organization that aims to attractive long-term power purchase agreement. bring solar power to owners of small buildings by reducing the cost per kWh

------111 nd 107 CAPEX is the capital expenditure or expenditure on the total cost of a power plant. Source: New York Times accessed at http://nyti.ms/PFfy on May 22 2013 112 nd 108 Ready to install kits which consists of all components required for installation. Source: CBS News accessed at http://cbsn.ws/1b73y6e on May 22 2013 113 109 Such a model is followed by e.g. SunEdison: www.sunedisonemea.com/installers/ accessed Source: 1 Block off the Grid accessed at www.about.1bog.org/faq/#eligible_cities on on May 22nd 2013 May 22nd 2013 110 For more information, see: http://about.1bog.org 114 Refer to the chapter on “Viability of rooftop solar PV” for details.

44 45 To make this model attractive for the consumers, the investor should offer a would not like the roof to be utilized for solar, the project developer might have solar tariff lower than the current grid electricity tariff. Equally, the escalation no other option but to shift the system to another location. Shifting an existing of the solar tariff should be lower than the expected escalation of the grid- system to another roof would create significant additional costs and would electricity tariff. push back the viability of solar (there is perhaps an attractive opportunity for a “smart”, easy to move mounting structure and balance of system design). The key concern in this model is the credit-worthiness of the power off-taker Another challenge to the model is a current lack of regulatory support for The key concern in (and hence the bankability of the project). To invest into a power plant on selling distributed power directly to end customers or utilities in Delhi. They key USP of this the RESCO model is the someone else’s premises and sell power exclusively to them under a 15 or 25 model is that it unlocks credit-worthiness of the year PPA requires a lot of faith in the customer. They key USP of this model is that it unlocks a greater number of residential a greater number of rooftops for PV systems, by providing economies of scale to the developers and power off-taker (and The key USP of this model is that the consumer gets electricity at a tariff an easy income opportunity to a rooftop owner. residential rooftops for hence the bankability of lower than the grid tariff without having to invest into the plant. The consumer, PV systems, by providing the project). moreover, has very limited risks (mostly related to potential leakages on the The city of Gandhinagar in Gujarat, has initiated a pilot project that has some economies of scale to the rooftops), no hassle and does not need to get involved in a new solar business characteristics of the solar rooftop leasing model described above. developers and an easy case. Hence, the investor should find it easy to convince consumers to go solar income opportunity to a and save without spending. Case Study rooftop owner. 8.3 local Micro Utility Model Gandhinagar solar rooftop program (2012 and on-going)

One way to reduce the off-take risk for a RESCO investor is to give solar power Gujarat has launched a pilot project in Gandhinagar to demonstrate the generators easy and cheap access to the distribution grid and allow them to working of grid connected rooftop solar power generation. The 5 MW project 115 sell power to third parties. Then, solar power developers could rent large, is based on a Private Public Partnership (PPP) model. If successful, it is to bundled roof spaces from building owners in a designated area, install PV then be replicated in five other cities of Gujarat. As per the PPP agreement, systems and sell the power generated to the rooftop owners, other consumers the Gujarat government has appointed the private project developers Azure or the DISCOMs at a pre-negotiated tariff. Having more off-take options would Sun Energy India Private Ltd. and Avantha Solar Power Maharashtra Private greatly reduce the risk to the project developer and improve the bankability of Ltd. to build, finance, own, operate and maintain grid connected rooftop solar the project. PV systems on approved rooftops of government buildings (4 MW) and private residences (1 MW)116. The developers feed the power into the grid at a pre- The project developers would particularly target those consumers who determined feed-in tariff of M11.21 ($0.22, €0.17) /kWh for Azure and M11.78 might not have the resources or would be unwilling to invest in rooftop solar. ($0.23, €0.18)/kWh for Avantha Solar. The PPA counterparty is Torrent Power, Developers can offer building owners a lease income on their rooftop space. the private utility that services Gandhinagar. Torrent Power fulfills its solar RPO obligation, enforced by the state of Gujarat through this agreement. The This model allows project developers to bundle rooftop space in a community project developer will pay a “green incentive” of M3 ($0.06, €0.05/kWh), as rent and thereby minimize the legal, commercial and technical transaction costs to the rooftop owners. The duration of the PPA is 25 years. The implementation by increasing the size of individual plants. This makes the model especially of this program was facilitated by the Gujarat Power Corporation Limited Through economies useful for the deployment of solar for residential consumers. The viable solar (GPCL). of scale, the cost per tariff in 2013 for a 50 kW plant is M8.6 (€0.13, $0.17) (close to the grid tariff kWh of solar power can for residential consumers) compared to the much higher viable solar tariff be significantly reduced of M11.60 (€0.18, $0.23) for a 5 kW plant. Through economies of scale, the and made viable for cost per kWh of solar power can be significantly reduced and made viable for residential rooftop owners (see also the earlier case study of 1 BOG). residential rooftop owners. Building owners will be able to generate new income through leasing their roofs for a period of time (e.g. 15 years). Renting homes is common in Delhi. The tenant might be disinterested in investing into a PV system that might have a payback period and a lifetime significantly exceeding the lease period. Such a leasing model would incentivize landlords who are renting out a building, to still make available their rooftops for PV plants.

Such a model can be combined with the RESCO model, wherein the project developer could sell power back to the building owner as well. In such a case, the lease for the rooftop might be waived in favor of a lower solar power tariff. ------115 PPPs broadly refer to long-term, contractual partnerships between the public and private sector agencies, specifically targeted towards financing, designing, implementing, and operating The key drawback in this model is the shifting of building owners from time to infrastructure facilities and services that were traditionally provided by the public sector time. In case the owner of the building sells the building and the new owner (Asian Development Bank 2006) 116 For more information, please see: http://bit.ly/16LUBTm . Source: The World Bank

46 47 Figure 20: The implementation structure of the Gandhinagar Solar Rooftop Figure 21: Gross metering Program

Source: International Finance Corporation (IFC), BRIDGE TO INDIA

The Gandhinagar rooftop pilot project will be launched in two phases of 2.5 MW each. The project is still at a nascent stage. Only 150 kW has been installed as on April 2013117.

The program makes rooftop PV for residences viable through the enforcement of the solar RPO requirement and the resulting willingness of the utility to pay a higher tariff. Given the higher grid tariffs in Delhi, such an approach is not needed: parity-driven solar will work with net-metering without additional Source: Gandhinagar Rooftop Program, BRIDGE TO INDIA incentives. However, the Gandhinagar rooftop model is a good pilot: The leasing model enables a rooftop owner to benefit from solar PV through earning rent in the form of the “green incentive”. The utility is able to meet its solar RPO and provides a bankable off-take for the PV system. Solar power generation is not limited by the load requirements of the building the plant is constructed on. Thus, available space can be optimally used. The tariff is, however, quite high The metering scheme adopted by the program is called “gross metering” (as and Torrent Power should be able to meet its solar RPOs in a cheaper way by opposed to “net metering”). The energy generated by the PV system is metered buying solar power from one of the many large, ground mounted solar plants separately from the consumption of electricity from the grid. As shown in in the state or by buying RECs. figure 2, “meter 1” measures the electricity consumed from the grid by the rooftop owner and “meter 2” accounts for the electricity generated by the PV system that is fed into the grid to be sold to the utility. At no point is the power from the PV system consumed by the rooftop owner. Since a FiT is offered as part of the program, a gross metering system makes more sense then a net metering system. The danger is that the two circuits might be fraudulently linked in such a way that grid power is sold back to the utility again through ------the solar meter, at a premium. Addressing this challenge will become more 117 Source: The Statesman accessed at http://bit.ly/185dwcg on May 10th 2013. important if the model is scaled. No official source available.

48 49 For commercial and industrial consumers, going solar can be a good way 9 why rooftop solar works of improving the “Corporate Social Responsibility” (CSR) footprint. It can for the stakeholders communicate a company’s concern about the community and the environment it operates in. Given that solar is viable, it is a CSR initiative that also makes economic sense. As an example, Shoppers Stop in Mumbai, a mall, installed 9.1 why rooftop solar works for a 30 kW solar system to promote itself as a “green” company and to hedge Delhi’s power consumers against rising costs of power in Mumbai120. Some global corporations that have already shown the way in other countries are Wal-Mart, IKEA and P&G. Since Figure 22: Benefits to consumers 2008, Wal-Mart has set up solar installations on 31 facilities in California and As a local, de-centralized Hawaii and has reduced energy costs by M50m (€0.76m, $1m)121. source of power, rooftop solar allows consumers As per the Companies Bill 2011122, companies with a profit of M5 billion (€0.07 billion, $0.1 billion) or more, a turnover of M10 billion (€0.15 billion, $0.2 to be more in control of billion) or more, or a net profit of M50m (€0.76m, $1m) or more, in a fiscal their electricity supply. year, are mandated to spend 2% of their net profit on CSR activities. “Ensuring environment sustainability” is one of the nine activities that qualify as a CSR initiative, according to the bill’s directives. The CSR obligation is likely to be enforced soon. Apart from being an obligation for larger corporations, CSR is increasingly being viewed by large and small businesses as a part of the overall business strategy. Going solar can provide industries with an additional “sustainability edge” in marketing their products in India and (perhaps more effectively) globally.

Currently, many power consumers in Delhi have stable power supply. However, it is not clear, if this will continue in the future. As a local, de-centralized source of power, rooftop solar allows consumers to be more in control of their electricity supply. Consumers can produce and consume their electricity on- site, making their electricity supply very reliable and secure. 9.2 why rooftop solar works for Delhi’s DISCOMS Source: BRIDGE TO INDIA Figure 23: Benefits to the DISCOMS Delhi’s power consumers are worried about the recent trend of rapidly Based on current trends, rising power prices. Solar can allow them to become more independent and, electricity tariffs can whenever parity is reached for a consumer segment, it can reduce their power 118 be expected to rise costs. Based on current trends , electricity tariffs can be expected to rise by a further 25% in the next five years for domestic (residential and government by a further 25% in consumers) and non-domestic (commercial consumers) segments and 29% for the next five years for the industrial segment in Delhi. domestic (residential and government consumers) On the other hand, solar prices are expected to decline at the long-term and non-domestic rate of 5-8% driven by technology improvements (e.g. on module efficiency and manufacturing improvements)119. Balance-of-system (BOS) costs (commercial consumers) (inverters, mounting structures, installation costs) will drive future plant cost segments and 29% for the reductions. Thus by switching to solar, residential, commercial, industrial industrial segment and governmental power consumers, can hedge against increasing electricity in Delhi. prices in Delhi. Also, if they are allowed to sell power to third parties or if a rooftop leasing model takes hold, solar offers a new income stream to rooftop owners in Delhi.

Source: BRIDGE TO INDIA ------120 nd 118 As explained in the chapter “Viability of rooftop solar PV” Source: The Economics Times accessed at http://bit.ly/OCarHB on May 22 2013 121 nd 119 The rate of decline of solar prices will reduce from the 50% decline seen in the past two years as Source: USA Today accessed at http://usat.ly/110x6OA on May 22 2013 the drastic module oversupply situation is easing and will be corrected in the next years globally. 122 Refer http://www.mca.gov.in/Ministry/pdf/The_Companies_Bill_2011.pdf for more details

50 51 DISCOMS should not view solar as an additional burden, but rather as a new 9.3 why rooftop solar works for business opportunity and as a chance to be at the leading edge of what utilities will likely be in the future: integrators of a smart grid with many different Delhi’s government DISCOMS should not sources of power generation and consumption. Fossil fuel supplies remain Figure 24: Benefits to the government view solar as an strained and carbon emissions need to be curtailed, while power demand 123 additional burden, but continues to rise . Significantly, rooftop solar PV is already financially viable for government consumers today and will be for commercial, industrial and rather as a new business residential consumers in the coming years. Adopting decentralized solar now opportunity and as a is a good opportunity for Delhi’s DISCOMS to begin preparing for the future. chance to be at the leading edge of what And more than that: utilities are ideally placed to capture a large part of this utilities will likely be in new market, for example, if they become part of the third party investment market (the RESCO or aggregator models described above). the future: integrators of a smart grid with In Delhi, often, the actual peak summer demand exceeds the planned, causing many different sources power outages for consumers. In addition, at times the power drawn from of power generation and other states fails due to inter-state grid problems, causing outages. Delhi’s consumption. grid infrastructure, too, collapses at times under the strain of excess electricity loads during peak summer months.

Given acute power shortages in Northern India, Delhi’s DISCOMS might be able to better plan loads and hence reduce the cost of buying power and maintaining the high level of power availability, if they explore alternatives to sourcing power from large, conventional plants located outside the city. In order for the utilities Localized, distributed renewable generation is a readily available answer. 2 GW to making the transition of rooftop solar PV would make a significant difference. Also, distributed solar Source: BRIDGE TO INDIA from an “old” to a “new” power, if implemented well, might help to stabilize the Delhi grid.

power system with In order for the utilities to making the transition from an “old” to a “new” Taking a progressive approach to the adoption of solar in Delhi could yield significant distributed power system with significant distributed solar PV, they will need clear political numerous advantages for the Delhi government. It would be a strategic choice solar PV, they will need leadership with a dependable roadmap and guidelines as well as support in to take an early and bold step into a new energy future. Solar can help Delhi increase its energy security and limit consumer power price rises. It could also clear political leadership building required infrastructure additions. Successfully implementing such a create an ecosystem for innovation, job-creation and learning to propel Delhi to Solar can help with a dependable transition in Delhi will put them in a very strong position to replicate the model in other markets. the forefront of a modern energy system. Delhi increase its roadmap and guidelines energy security and limit as well as support Delhi’s peak power demand has increased from 3,097 MW in 2002 to 5,642 124 consumer power in building required MW in 2012 . While the power demand has increased, the power generation capacity within the state has not kept pace. In 2002, Delhi had a power price rises. infrastructure additions. generation capacity of 527 MW. This has increased to only 1,344.7 MW by 2012, supplying only 28% of the city’s demand125. Delhi’s demand is expected to reach over 11,000 MW in 2022126. At the moment, the only expected additions in power generation capacity in the city are a 1,370 MW gas based plant and a 94.8 MW coal based power plant. Land in Delhi is expensive and unavailable to set-up large conventional power plants.

A large part of Delhi’s peak power demand is met by purchasing power from generators in other states through long-term PPAs (meeting 58% of Delhi’s demand) and purchasing short-term power through power exchanges and

------124 BRIDGE TO INDIA analysis based on CEA data. www.cea.nic.in/more_upload/epsr_17_ highlights.pdf accessed on May 22nd 2013 125 Source: DERC tariff orders. ------126 BRIDGE TO INDIA analysis based on CEA data. www.cea.nic.in/more_upload/epsr_17_ 123 Refer http://bridgetoindia.com/blog/?p=1638 for more details highlights.pdf accessed on May 22nd 2013

52 53 from generators in other states (meeting 13% of Delhi’s demand). Relying Adoption of solar would also be accompanied by the creation of new on external sources, however, has made the city’s power supply vulnerable. employment opportunities for the city. The solar industry is capable of 82% of Delhi’s power from external long-term PPAs is from gas and coal. developing an ecosystem that creates jobs for system installers, transporters, Gas supplies are short127. Increased supply of large quantities of gas will project developers, system maintainers, grid monitors and power quality only be possible with new pipelines that might not be built, and even if they controllers. The solar industry has the capacity to create 32 job-years/MW135 are, will take time. In addition, there is a shortage of coal supplies in India. that implies that 62,000 new full-time jobs can be created by installing 2 GW of Coupled with water shortages, this leads to a fall in generation in coal plants. solar power. With an annual influx of 70,000-80,000136 migrants to Delhi, and In December 2012, the northern region faced a power deficit of 8.2% owing to an unemployment rate of 4.8%137, these new jobs could absorb a considerable By focusing on energy such shortages128. Delhi was directly affected by such shortages in October amount of the workforce. security, cleaner energy 2011, when production of coal plants supplying power to Delhi from outside the and job creation, the Delhi state fell by over 1,000 MW, leading to severe power cuts in the city129. The city’s By focusing on energy security, cleaner energy and job creation, the Delhi Rooftop solar is the DISCOMS partly compensate for such drops in generation by overdrawing from government can project itself as a progressive administration. The aggressive government can project only locally available the northern grid, purchasing short-term power from other states and the spot adoption of solar can enable Delhi to take up the role of a green leader and set itself as a progressive power generation market. an example for other states as it did in the case of Compressed Natural Gas administration. option available to (CNG). The city of Delhi also has the capability of becoming the center for solar Purchasing short-term power from the northern grid presents a serious risk research and development or an “innovation hub” for the country. For this, it the government. to the city’s power supply. In the financial year 2012-13, the northern grid can exploit the knowledge pool of top universities such as the Indian Institute faced a peak deficit of 12.3%130. This summer, states like Uttar Pradesh will of Technology (IIT), located in the city. lack almost 7,000 MW and Punjab 3,000 MW131. Such states will draw on power from Rajasthan, Jammu and Kashmir (facing its own power deficit of 50%) and Gujarat. Delhi, too, relies on these and other states for short-term power, sources that will be strained and increasingly unavailable in a scenario of continuing power deficits. The reliance of Delhi’s DISCOMS on power from the northern grid only further exposes the city to the kind of massive grid failure that took place in July 2012132. Rooftop solar is the only locally available power generation option available to the government.

As per the World Health Organization’s (WHO) data, covering 1,000 cities in 91 countries133, Delhi is the 12th most polluted city in the world. The respirable suspended particulate matter (RSPM) and the concentration of Nitrogen Oxide (NOx) are above permissible levels134. This in turn has led to an increase in the number of respiratory disorders and cases of hypertension. It is important for the government to make a gradual shift to a cleaner source like solar to improve the health parameters of its residents. Solar could give the city the room for manoeuver that it needs to shut down the heavily polluting, coal fired power plants within the city.

------127 Refer http://bit.ly/18QC05L accessed on May 22nd 2013 128 Source: CEA 129 Source: The Hindu accessed at http://bit.ly/19KI8xR on May 22nd 2013 130 Source: CEA Load Generation Balance Report 2012-13 accessed at http://bit.ly/11GsqMv on May 22nd 2013 131 Source: Ministry of Power ------132 Refer The New York Times article at http://nyti.ms/Mwq7Xj accessed on May 22nd 2013 135 As per the reports “The work that goes into renewable energy” by Renewable Energy Policy 133 Using air monitoring data from 2003-2010 Source: Slate accessed at http://slate.me/W2zWmT Project -2001 and “Putting renewables to work” by University of California, Berkley -2004 around on May 22nd 2013 5 jobs (excluding manufacturing of modules, inverters and cables) can be created by 1MWa of solar. With 16% efficiency, 32 job-years can be created by 1MW of solar installed. 134 RSPM in the Capital's air is touching 250 micrograms per cubic meter (ig/ m3) when Refer http://bit.ly/19KJby2 and http://bit.ly/nCvEne accessed on May 22nd 2013 the permissible level is 60 micrograms per cubic meter and concentration of nitrogen oxide (NOx) 136 nd is 50-55 ig/ m3 well above the permissible upper limit of 40ig/ m3. Source: India Today accessed Economic survey of Delhi 2012-2013 accessed at http://bit.ly/11GtcJl on May 22 2013 at http://bit.ly/Ngk0YM on May 22nd 2013. 137 Source: http://bit.ly/10Sg4pg accessed on May 22nd 2013

54 55 10 Delhi solar roadmap: Figure 25: Priority scenario expected annual solar capacity additions (in MW) reaching 2 GW by 2020

Delhi can become a leading solar city in the world without any significant investment into either incentives or the grid, based on the commercial viability of solar power and the existing infrastructure.

What is required to reach that goal, however, is a fundamental shift in the way that consumers, DISCOMS and regulators think about the provisioning of power. Distributed solar power should not be regarded as a nuisance, as We have outlined being irrelevant, a regulatory challenge or cost factor. It is neither of these. It is two scenarios for the rather a viable, achievable, significant part or the solution to Delhi’s long-term implementation of 2 GW power requirements. by 2020. The priority We have outlined two scenarios for the implementation of 2 GW by 2020. The scenario is based only on priority scenario is based only on grid parity and does not involve any incentives grid parity and does not by the government. The optional scenario is based on an aggressive approach involve any incentives with significant capacity addition even before parity is reached. This would by the government. The require incentives from the government and would make sense only, if there is an acute concern about energy security. optional scenario is based on an aggressive 10.1 Priority scenario (parity based) approach with significant capacity addition even In this scenario, the solar installations are purely based on commercial viability before parity is reached. once parity is reached for a consumer category. Solar is already viable for government consumers today, if projects are bundled and tendered together. For commercial type-1 and industrial type-2 consumers, solar will be viable in 2014 and they should be ready to switch to it. In the course of the following five years, the other three power consumer categories (commercial type-2, industrial type-1 and residential) also reach parity.

We foresee two major growth periods: one between 2014 and 2017 when the government buildings go solar and one between 2018 and 2020, when residential installations will come into the market. The graph, on the facing page, shows the expected annual solar capacity addition per tariff category.

In predicting the growth rate, we have made the following assumptions: 1. Growth will be driven by parity only (and not by incentives). Therefore, only once parity has been reached for a tariff group will solar be installed. 2. We assume annual “adoption rates”. Once parity has been reached, it will still take time until all customers are ready to go solar. This lag might Source: BRIDGE TO INDIA analysis be related to e.g. unavailability of financing options, uncertainty about building ownership/use or a lack of awareness. 3. We assume a relatively low initial adoption rate (of e.g. 5% for commercial type-1 consumers and 5% for industrial type-2 consumers) that will finally max out at 80%, assuming that there will always be a residual group that will not adopt solar (an addition to the other discount factors included in the calculation of the potential). 4. Because the residential segment reaches parity late, it does not achieve 5. Our approach is slightly different for the government sector. We assume a 80% by 2020. 1% adoption in 2013, because we suggest that the government should do a trial tender in 2013 to learn the ropes. Following that, the adoption rate of the government is fast, because we assume that a transition to solar power can be managed through a limited number of actors. For the same reason, we assume that 100% of the potential can be converted by 2017.

56 57 Figure 26: Priority scenario cumulative solar rooftop growth until 2020 Figure 27: Optional scenario expected solar capacity addition (in MW)

Source: BRIDGE TO INDIA analysis

10.2 oPtional scenario (government incentives)

If the government wants to address energy security challenges more urgently through solar, it will have to provide a “viability nudge” through incentivizing the adoption of rooftop solar. This will lead to parts of the potential of rooftop solar being realized even before parity is reached. Once parity for a particular consumer category is reached, no further incentives are needed. Due to incentives, solar will be adopted by all the categories of consumers by the year If the government 2014. The installations reach a peak in 2017 and by 2018 all the categories, wants to address energy except residential consumers, reach their potential. The graph on the facing security challenges more page shows the expected annual solar capacity addition.

urgently through solar, In predicting the growth rate, we have made the following assumptions: it will have to provide a 1. The government will provide incentives to all the consumers that are yet “viability nudge” through to reach parity such that they can achieve a 15% internal rate of return incentivizing the adoption on their investments (or a 7 year payback for residential consumers). The of rooftop solar. incentives will reduce every year to factor the increase in grid tariffs and reduction in solar costs over time. 2. As in the priority scenario, industrial and commercial consumers will Source: BRIDGE TO INDIA analysis adopt 80% of their potential. Residential consumers too will adopt up to 80%, unlike in the priority scenario where they only reach 60%. This is because they will adopt early due to a viability nudge, and will adopt more of their potential over a longer period of time than other consumers. 3. The initial adoption rate is 5% across commercial and industrial consumers as there is lack of awareness and the market for these consumers will take time to strengthen. These adoption rates will makes out at 80% assuming that there will always be a residual group that will not adopt solar (an addition to the other discount factors included in the calculation of the potential). 4. As in the main scenario, the government will initiate adoption in 2013 by performing a small, trial tender for government buildings. The adoption in government projects will be similar to that in the parity scenario.

58 59 Figure 28: Optional scenario cumulative solar rooftop growth until 2020 There are a range of factors which could advance or delay the adoption of solar as per our roadmap.

Table 6: Factors impacting rooftop solar adoption

Faster adoption factors Slower adoption factors Availability of low cost debt Lack of financing or increases in would make solar more viable. the cost of debt would reduce the Standardized financing products viability of solar. e.g. for the residential market would raise the adoption rates.

A fall in the supply security of grid Legal uncertainties and other power would result in more on diesel complications around ownership back-up usage and a higher average of rooftops would slow down the power cost138. This would bring parity adoption of solar and reduce the

Source: BRIDGE TO INDIA analysis forward. overall adoption rate.

Grid tariffs rising faster than 5% p.a. If government customers receive or solar costs falling faster than 6% subsidized or free power, their Delhi would be able to reach over 2,000 MW by 2020 (2,160 MW in the priority p.a. would advance parity. incentive to go solar will fall scenario and 2,284 MW in the optional scenario). This amount of peak power dramatically. generation can be grid interactive. The grid should be able to accommodate this amount of solar power, while remaining stable. The graph below shows If market aggregators can combine Delays in metering policies and the cumulative amount of solar power in the grid and the ability of the grid projects e.g. in the residential, connectivity standards would reduce to handle intermittent power (at 20% of total capacity). In both the scenarios, commercial or industrial segment, adoption rates especially in the the capacity addition is not crossing the grid’s solar handling capacity. In the they can make use of economies of residential segment, would reduce alternative scenario, the capacity addition touches the grid handling capacity, scale and reduce solar costs average system sizes and would hinting that further capacity addition would require upgrades in the grid. push back parity for the industrial and residential segment. Figure 29: Grid handling capacity vs. expected solar growth

Source: BRIDGE TO INDIA

Source: BRIDGE TO INDIA analysis ------138 Diesel rates have increased by over 259% in the last 10 years. The per kWh cost of diesel is over M14 (€0.21, $0.28).

60 61 11 government policy Conduct stakeholder meetings and more detailed analysis recommendations The government should bring together the affected stakeholders of such a power shift to solar in regular stakeholder meetings. These would include the In our priority scenario we suggest a parity-driven rollout of Delhi’s solar various regulatory bodies (DERC, CEA, etc.), the Delhi utilities, the utilities of potential. In this scenario, the government of Delhi would not have to provide the adjoining states Haryana and UP (who will be affected, too), civil society incentives to improve the viability of solar. However, the role of the government representatives (e.g. Greenpeace, academic and research institutes, Resident would still be crucial. It has to create a level playing field with respect to Welfare Associations, other customer groups), solution providing companies The DERC should grid-usage and enable the growth of an eco-system that can deliver quality and financiers. In addition and perhaps steered by the stakeholder meetings, interact with the PV systems at competitive prices. If the government opts for the optional, the government should commission a detailed analysis of the technical state DISCOMS, accelerated scenario involving incentives, its role will become even larger. challenges and solutions of interconnecting large amounts of PV with Delhi’s grid. This would prepare Delhi for the later stages of PV development (2016 system installers and Figure 30: Policy recommendations for the government onwards), but such a study should not be a pretext for slowing-down or independent consultants stopping initial PV growth. to frame Delhi specific standard and guidelines Introduce regulations for PV components and metering (gross or net) in-sync with the ground Metering options as well as standards and guidelines for rooftop PV systems realities, challenges are already being developed by the CEA in a forthcoming report on grid and motivations of the integration of solar PV (currently under review). However, the onus of selecting stakeholders. the metering policy (gross or net) for the state of Delhi would fall on the Delhi Electricity Regulatory Commission (DERC). The DERC should interact with the state DISCOMS, system installers and independent consultants to frame Delhi specific standard and guidelines in-sync with the ground realities, challenges and motivations of the stakeholders. Further, guidelines specifying technical standards of connection, procedures of applying for such a connection and configuration of meters and their arrangement also need to be formulated.

Strengthen load forecasting by DISCOMS

As per the roadmap for solar adoption in this report, 2 GW of solar by 2020 will be within the limits of the grid. However, the impact of PV on the grid in Delhi is not entirely understood. There is a need for government policy to prepare Source: BRIDGE TO INDIA DISCOMS for the expected ramp-up of PV capacity. Government policy needs to ensure that rooftop PV’s contribution is included in the power planning process of the DISCOMS. DISCOMS need to include solar into the city’s energy Government policy growth forecasts. The government can support DISCOMs in developing more needs to ensure that sophisticated weather forecasting, which would enable DISCOMs to better manage solar power scheduling. rooftop PV’s contribution is included in the power Support the managing of solar power intermittency through planning process of the gas-based power DISCOMS. 11.1 Recommendations for enabling Delhi today has a little over 600 MW of gas based power generation (owned parity based solar growth by the Delhi government). Another plant of 1,370 MW by Indraprastha Power Generation Company Ltd. (also owned by the Delhi government) is planned and 1. Provide support for grid-connectivity of rooftop solar could be operational in 2013. Gas based power plants can ramp up quickly and can balance the intermittency of solar. The costs of using gas power flexibly, Owners of rooftop PV systems will want to feed power into the grid during however, would increase the per kWh cost of the gas power (as the plant is not those times when their captive power consumption is low (e.g. on public optimally used). The government might consider stepping in with an incentive holidays and weekends). Feeding power into the grid would allow for a more to make viable the use of gas-based power as a hedge against solar power generous system sizing and would significantly reduce the per kWh cost intermittency. of solar as no power is wasted – thus making it viable more quickly. The government would need to facilitate metering (either gross or net).

62 63 Simplify installation guidelines and standardize permitting and including the City University of New York (CUNY), the Mayor’s Office of Long- connectivity procedures Term Planning and Sustainability and the NYC Economic Development Corporation. The U. S. Department of Energy (DOE) provided funding and the Rooftop solar is new for government administrators as well the city’s National Renewable Energy Laboratory (NREL) provided technical support. DISCOMS. It is important that rooftop solar specific guidelines are incorporated Work on the New York solar map began in 2010 and took about 14 months in to Delhi’s building codes. The government should also set standards for the to complete. The overall cost of setting up the solar map was around M33m quality of installations and the structural integrity of buildings where rooftop (€660,000; $858,000)139. It is important solar is to be installed. Guidelines need to be clear and simple so as to allow for the government consumers to meet them and officials to enforce them easily. They should also The New York solar map enables homeowners to determine the potential of to communicate the be made available on an online portal for transparent and easy access to all solar on their buildings and assess the financial viability of installing a solar potential, the viability the stakeholders. This will help limit the complexity and time taken for setting system. The map also shows the existing installations, which allows installers up rooftop installations. to assess trends in adoption of solar and help them better target their business and the opportunities efforts. Further, utilities can use the information to estimate the amount of of rooftop solar to all 2. Communicate the potential and feasibility of customers switching to solar energy. Finally, state planners and government the stakeholders in rooftop solar bodies can track the progress towards achieving targets such as New York’s the city. Renewable Portfolio Standard. Increase public awareness San Francisco was the first city in the US to develop a solar map in 2008. After the map’s launch, PV installations in the city grew by 60% in two years140. Since It is important for the government to communicate the potential, the viability then, 18 cities in the United States have followed suit to share information and the opportunities of rooftop solar to all the stakeholders in the city: to related to solar with their residents. consumers, DISCOMS and government administrators. The government could articulate an official Delhi Solar 2020 vision document with a roadmap for Delhi Figure 31: Screenshots of the New York solar map to become a leading solar city globally.

Solar is still widely perceived to be an expensive source of power and there is too little knowledge about the amount of power a PV system can provide and where it can be procured. There is a need to demystify solar costs and prove to the stakeholders that it is viable. Consumer’s electricity bills are a good platform to relate electricity prices with solar costs to show their viability. The government could mandate Delhi’s DISCOMS to include viability analyses on A comprehensive, consumers’ power bills to raise awareness about solar. online tool that maps Delhi and offers a Provide an online “Delhi Rooftop Solar Map” calculation for the A comprehensive, online tool that maps Delhi and offers a calculation for the benefits of solar would go benefits of solar would go a long way in encouraging especially the residential a long way in encouraging power consumers to go solar. Ideally, this would be a very user-friendly tool, especially the residential based on a geographical information system (GIS) mapping of rooftop PV power consumers space in Delhi. Then, a solar calculator would offer building owners a simple to go solar. viability analysis based on grid power costs and available rooftop space. Such a calculator already exists for India on the website www.indiasolarhomes.com. This could be further detailed for Delhi. New York, for example, has such a map in place for its residents.

Case Study

The New York solar map

The New York solar map was introduced in 2011 to provide clarity on government policy and the potential for solar PV to assist all the stakeholders. The map provides quantitative data on installation costs, cost savings, energy savings, and local, state and federal incentives.

The map was initiated by the New York Solar America City Partnership ------139 Source: American Planning Association accessed at http://bit.ly/UYKi6E on May 22nd 2013 140 Source: American Planning Association accessed at http://bit.ly/UYKi6E on May 22nd 2013

64 65 Case Study

German House solar rooftop plant, Delhi

Source: American Planning Association NYC Solar Map141

Set-up demonstration projects on transport facilities in Delhi

At the moment, solar PV is understood primarily as a large MW, grid- connected source of power. There is a need to make the idea of rooftop solar in Delhi real for all the stakeholders through inspiring pilot projects. Transport There is a need facilities like railway stations and the Delhi airport have a low overall potential to make the idea of for rooftop solar as per our analysis (2.6 MW). Hence, we have not included them in the target and roadmap. However, they are excellently placed as rooftop solar in Delhi real Photograph by GIZ ComSolar demonstration sites with very high visibility. for all the stakeholders The German House in New Delhi has installed two solar PV systems with a through inspiring The Indian Railways has a CSR obligation as per recent guidelines issued by total system capacity of 10.5 kWp on its rooftop in 2011. Q.cells, a German the Heavy Industries and Public Enterprises Ministry142. The Delhi Airport, too, pilot projects. module manufacturer (now owned by Hanwha), financed and installed the has a CSR mandate (though not enforceable) as per guidelines by the Ministry systems as a demonstration project to showcase the technical feasibility of a of Corporate Affairs. The Delhi Airport is already planning to build a 2 MW, PV rooftop system in Delhi. For comparative reasons, the two solar PV systems ground-mounted solar power plant near the runways. The German House in use different technologies – thin film and polycrystalline modules. New Delhi has set up a demonstration plant on its rooftop showcasing different technologies. The plant was commissioned in August 2011 and can provide a The solar power generated from the plant is connected with the internal good body of actual generation data. energy management system of the building, where it is combined with electricity from the grid. The PV installation system uses a remote, wireless monitoring system, which transfers and stores generation data. The panels are cleaned thrice a week.

Table 7: German House solar project details

Total system capacity 10.47 kW Commissioning date 19.08.2011 Module technology Polycrystalline (5.52 kW) and thin film (4.95 kW) Module area 40 m2 (polycrystalline) and 41 m2 (thin film) Inverters 2 x SMA SMC 6000 ------Modules 24 x Q-CELLS Q.Base 230 and 141 nd Accessed at www./nycsolarmap.com/ on May 22 2013 55 x Q-CELLS Q.Smart UF 90 142 Source: iGovernment article accessed at http://bit.ly/19KKtcg on May 22nd 2013

66 67 Storage of electricity None There is a strong business case for third parties to invest into the PV plants of such consumers, offering solar power as a service, with no investment cost Tracking None to the consumer and immediate savings on their electricity bills. However, Annual electricity generation 15,097 kWh identifying viable customers can be difficult for investors. The government Monthly electricity generation 1,258 kWh could launch a Cost of system M2,000,000 (€30,789, $40,000) The government could launch a targeted program of workshops offering targeted program of Average annual CUF 16% a free assessment of the techno-economic feasibility of solar for building owners. Such a program would benefit customers that are still unsure about workshops offering Source: GIZ ComSolar the viability of solar. The government could also invite interested investors or a free assessment of EPC companies to participate in these workshops to propose their solutions. Figure 32: Electricity generation and CUF of PV plant at the German House, the techno-economic A similar program has been successfully executed in San Francisco. New Delhi feasibility of solar for Case Study building owners.

The Mayor’s Solar Founders’ Circle Program in San Francisco, US (2008-2009)

In 2007, the U.S. Department of Energy (DOE) designated the city of San Francisco in California a “Solar America City”. The city has since been aggressively promoting rooftop solar through incentives and dedicated programs. One such program was “The Mayor’s Solar Founders’ Circle”143, launched in 2008. It targeted 1,500 commercial rooftop owners in the city for the installation of solar PV systems. The program offered a free personalized technical and financial analysis of the solar potential of the buildings. More than 100 companies applied for the energy audit, out of which 40 were audited144.

The City of San Francisco partnered with the National Renewable Energy Laboratory (NREL) to train city employees and consultants to perform site assessments. A team of national lab experts gathered data and formulated individual reports. Large scale solar 3. Create a knowledge and skills base for executing solar deployment will be successful only if there Large scale solar deployment will be successful only if there is a strong is a strong focus on focus on ensuring quality and an adherence to technical standards in the ensuring quality and an PV installations. This requires a well-developed skills and knowledge base adherence to technical in the city. Delhi is in a unique position to develop this base as it has access to top engineering universities like the Indian Institute of Technology and standards in the numerous technical institutes and degree colleges. The government should PV installations. initiate a program across such institutions to include technical courses on PV installations. The technical assessment and monitoring of government Source: GIZ ComSolar projects should be handed over to educational institutions so that students in PV courses practically engage with installations in the city and form the first cohort of technicians needed to support large-scale solar deployment in Conduct solar feasibility workshops for commercial and industrial the city. consumers

While rooftop solar will be viable for commercial type-1 and industrial type-2 consumers next year, it is possible that such consumers do not have enough information on the economic case for solar, are deterred by high upfront investment costs, are uncertain about the technology or worry about potential ------structural risks (e.g. for leakages in the roof). 143 Source: Office of the Mayor, City and County of San Francisco accessed at http://bit.ly/13mZkG9 on May 22nd 2013 144 Source: US Department of Energy, Energy Efficiency and Renewable Energy accessed at http://1.usa.gov/17qEcTE on May 22nd 2013

68 69 4. Prepare projects for government tenders power, it might decide that it wants to drive a more aggressive adoption of rooftop solar. Residential consumers have the largest potential (1.2 GW), but, Bundle government projects in a parity-based scenario will begin contributing to Delhi’s power supply only in the year 2018. Similarly, commercial and industrial consumers only begin Viability nudges from Solar for government consumers in Delhi is already viable in 2013, as long as in 2014 with certain types of consumers going solar as late as 2016. The key the government can be in projects are bundled to a cumulative size of around 2 MW. Bundling provides to activating the early adoption of solar by such consumers without waiting for the form of a generation parity is for the government to make solar artificially more attractive, providing scale to the installer who would save on engineering, procurement and based incentive (GBI) or The Delhi government logistics costs. The Delhi government should aggregate the buildings it owns a viability nudge. capital subsidies. should aggregate the and tender these out to RESCOs that want to invest in rooftop PV and provide Viability nudges from the government can be in the form of a generation based buildings it owns and the government with solar power. A tender will be a competitive process, the benefit of which will be directly passed on to the government exchequer. incentive (GBI) or capital subsidies (these can be further supported with tax tender these out to breaks or low cost debt). We calculated the cost of both, a GBI and capital RESCOs that want to Learn the ropes through demonstration projects in 2013 subsidies, to the government to achieve its target of 2 GW by 2020145. As shown invest in rooftop PV and in the graph below, a GBI could cost the government a total of M2.5 billion provide the government The government could release a small tender already in 2013. In our roadmap, (€0.04 billion, $0.05 billion) by 2023. On the other hand, capital subsidies could cost a total of M5.8 billion (€0.08 billion, $0.11 billion) by 2018, twice the cost of A GBI could cost with solar power. we suggest three projects of 2 MW each at three different locations in Delhi. These projects will serve as a test-case for larger tenders in the upcoming a GBI. the government a total of years and as a demonstration to central and state government administrators M2.5 billion (€0.04 billion, Figure 33: Cost to the government for viability nudges of the viability of solar for government buildings in Delhi. $0.05 billion) by 2023. On the other hand, capital 5. Improve the solar financing framework subsidies could cost a In order to overcome the liquidity challenge of solar power investments, debt total of M5.8 billion (€0.08 plays a crucial role. Debt at rates lower than the equity IRR expectations (15%) billion, $0.11 billion) by would also reduce the cost of solar on a per kWh basis, making it viable more 2018, twice the cost of quickly. In our viability analysis, we have assumed a debt to equity ratio of a GBI. 70:30 for government, industrial and commercial projects. Residential systems we assume to be purely equity funded.

If the cost of debt for the first three projects can be brought down (from a current 13% to say, 10%) and the lending risk be cushioned to induce more Debt at rates lower non-recourse loans, the adoption of solar could be significantly accelerated. than the equity IRR expectations (15%) would For the government, industrial and commercial projects, the government could provide a payment guarantee. also reduce the cost of solar on a per kWh basis, The government could consider setting up a risk guarantee fund for banks making it viable to lend more readily. This is being done, for example, by the Solar Energy more quickly. Corporation of India (SECI) for the National Solar Mission.

Banks currently place solar in their power sector/infrastructure lending portfolios. In order to unlock the residential sector, it would be crucial that banks offer consumer finance solutions for solar home systems (such as, for example, equal monthly installments (EMIs)). The Delhi government could support banks in developing such solutions through making available Source: BRIDGE TO INDIA analysis expertise, reacting to the risk concerns of banks, or even through setting up an own, dedicated non-banking financial company (NBFC). 11.2 oPtions for activating early adoption through a “viability nudge” ------143 Source: Office of the Mayor, City and County of San Francisco accessed at http://bit.ly/13mZkG9 At the moment, Delhi has access to 24/7 power. But, as we explain in the on May 22nd 2013 chapter on making rooftop PV work for the stakeholders, there are risks to 144 Source: US Department of Energy, Energy Efficiency and Renewable Energy accessed at http://1.usa.gov/17qEcTE on May 22nd 2013 Delhi’s power supply. For the government to maintain its promise of 24/7 145 We calculated incentives required for each consumer type to achieve a 15% IRR. We have assumed a 5% annual escalation in grid tariff and 6% annual reduction in costs of solar.

70 71 1. Viability nudge through a generation-based incentive competitive solar PV installation costs in the world149. With the deployment of (GBI) 30 GW150 in the past 13 years, benefits of scale have driven costs down – on a global level (e.g. for modules), but also for the local installation work. Further, as the country has moved along its learning curve it has created a robust Increasing the revenue from a PV system through a GBI would improve solar ecosystem for the German solar market with simple application procedures, viability. GBI would bridge the gap between the grid price and the viable easy availability of information, healthy market competition, effective industrial solar tariff for a consumer. In this approach, the viability nudge provided associations like the “Bundesverband Solarwirtschaft”151, strong research to a consumer falls over time as the grid prices rise. Finally, once parity is institutions such as the Fraunhofer ISE152 and experienced government reached, the expenditure to the government becomes zero. Thus, the GBI for bodies. Transaction costs related to e.g. site inspections, permitting, customer installations set up in 2014 would be phased out in 2024, when the grid tariff acquisition and financing are among the lowest in the world153. The eco-system reaches the viable solar cost of a solar plant constructed in 2014 (at M10.90/ The total expenditure has become so strong and predictable that banks offered inexpensive loans of kWh, €0.16/kWh, $0.21/kWh). In addition, the GBI offered by the government up to 100%154 of the solar rooftop system cost, making the adoption of solar a on GBI reaches a peak to new consumers would drop each year, as solar prices fall and grid tariffs “no investment” income opportunity for rooftop owners. in 2017, as new capacity increase. The GBI would be phased out for new installations once solar is dependent on a viability competitive with residential grid prices, expected by 2018 (at M8.51/kWh, nudge will be installed €0.13/kWh, $0.16/kWh)146. The total expenditure on GBI reaches a peak in 2017, 2. Viability nudge through a capital subsidy only until that year (as as new capacity dependent on a viability nudge will be installed only until that year (as residential consumers reach parity the last, in 2018). Another option for a viability nudge is to provide a direct, one-time capital The government’s residential consumers subsidy at the time of installation of the plant. For viability in 2014, residential expenditure on subsidies reach parity the last, In Germany, similar generation based viability nudge (through a flat FiT that consumers require a capital subsidy of 38%, type-1 industrial consumers is limited to five years in 2018). covers the entire viable solar tariff and not just the gap between the grid price require 25% and type-2 commercial consumers require 7% as per system (2014-2018), until all the costs in 2014. and the viable solar tariff for a consumer) has been hugely successful for consumer categories driving rooftop solar PV installations. In the case of capital subsidies, the government would bridge the entire reach parity. Case Study viability gap, upfront, at the time of installation. The government would effectively reduce the viable solar tariff to match the grid tariff by providing a Germany launched its first Feed-in Tariff (FiT) law, the subsidy. Since the full viability nudge is provided at the beginning of a project, “Stromeinspeisungsgesetz” (StrEG), in 1999. This bill required utilities to the government will be unable to take advantage of the reducing gap between connect renewable energy generators to the grid and buy the electricity the viable solar tariff and the grid tariff over subsequent years. This is the key produced by renewables at 65-90% above the average retail price of electricity. reason that subsidies could cost the government twice the amount required By 2004, Germany announced a nationwide FiT of M30-40147 (€0.46-0.62; $0.35- for GBI. In the graph above, the government’s expenditure on subsidies is 0.47)/kWh, independent of electricity prices, with an annual reduction of 5%. limited to five years (2014-2018), until all the consumer categories reach Tariff rates were to be revised every four years. In 2012, flexible tariff reduction parity. The expenditure will be the highest in 2016 as subsidy dependent PV was introduced, with a revision every quarter148. If the overall, newly installed capacity addition peaks in that year (driven by residential consumers and some A one-time, upfront capacity in the preceding 12 months was within the target corridor of 2.5-3.5 industrial type-1 consumers that will install PV before parity is reached for disbursement is GW, the tariff would be reduced by 1%. This rate increases, if the capacity them mid-year). administratively addition exceeds the upper limit of the corridor and decreases, if it falls short. convenient for the The flexible reduction structure is transparent and responds to developments A one-time, upfront disbursement is administratively convenient for the in the market. government. This however, does not justify the significantly higher cost than a government. This GBI. In the case of a capital subsidy, the government will also need to ensure however, does not justify The tariff structure incentivizes a potential consumer to go solar as soon as that consumers do not compromise on the quality of the PV systems and take the significantly higher possible in order to benefit from higher incentives. The tariffs for a project are undue advantage of the subsidy by inflating on-paper system costs but keeping cost than a GBI. fixed for a period of 20 years, making investment returns predictable, thus actual system costs extremely low. This could be achieved by setting standards reducing risks and financing costs. for installation and contracting independent monitoring agencies to assess their implementation. Even though the German government is now in the process of phasing out the tariffs, the market has gained significant momentum, with perhaps the most

------149 The average costs of PV systems in Germany in the second quarter of 2012 was M110/Wp ($2.2; ------€1.69) and in US is M275/Wp ($5.5; €4.23). Source: IRENA accessed at http://bit.ly/ZP2q6I on 146 Refer to Chapter 6 on “Viability of rooftop solar PV” May 22nd 2013 147 Differentiated by system size and application type (roof-mounted or ground mounted) 150 Source: Fraunhofer Institute accessed at http://bit.ly/SwvvdZ on May 22nd 2013 148 Under the amended EEG 2012, Germany will continue to adjust tariff rates periodically in an 151 Source: Bundesverband Solarwirtschaft attempt to keep annual PV installations within a target “corridor.” The current annual installation 152 Source: Fraunhofer Institute corridor is 2,500 to 3,500 MW. Also see : “The German Feed-in Tariff for PV”by Deutsche Bank 153 Source: Bundesverband Solarwirtschaft accessed athttp://bit.ly/Od0gDu on May 22nd 2013. Climate Change Advisors- May 2011accessed at http://bit.ly/11tceCt on May 22nd 2013 and See also: “Why Solar Installations Cost more in the U.S. than in Germany” Dec , 2012 “The German Feed-in tariff: Recent Policy Changes “ by Deutsche Bank September 2012 http://bit.ly/10haaOp accessed at http://bit.ly/1303qmc on May 22nd 2013 154 Press Release by kfW : http://bit.ly/15Mk2QS accessed on May 22nd 2013

72 73 Photograph by Anand Pathanjali/Greenpeace India Photograph by Anand Pathanjali/Greenpeace India

74 75 of plot area. Only built structures that were of sturdy concrete constructions 12 APPENDIX were included in the computation of the potential rooftop space, such that they could bear the weight of the solar rooftop installations. Finally, we discounted 12.1 methodology for calculating the for space that was obstructed with objects and structures and space that was solar rooftop potential in Delhi shadowed to determine the solar suitable rooftop area that receives optimal sunlight for solar power generation. In order to determine the geographic potential of Delhi to generate solar power, we have assessed the rooftop space offered by its built area. GIS is the 12.2 computation of the solar best technology available to get near accurate measurements of area through suitable rooftop space for different satellite mapping. On the downside, it costs around M60m (€0.92m, $1.2m) and its application takes a period of almost 9-12 months, considering that Delhi is land areas twice the size of Singapore. As we were working within a reasonable timeline of 2 months, the application of GIS was unfeasible for our analysis. Our analysis 12.2.1 Residential buildings is thus based on information provided by the Delhi Development Authority (DDA), a government-run organization that has mapped and measured the The residential area in Delhi is heterogeneous in nature and reflects the entire city of Delhi as part of its land development and planning practice. DDA’s city’s rapid, often uncontrolled growth. There are different types of buildings, Master Plan 2021 and Zonal Plans are the only official documents that provide ranging from individual houses, apartments, residential complexes, comprehensive information on different land areas with their designated uses unauthorized colonies, Jhugi Jhompri (JJ) clusters157 to slums. Further, in the future. these housing settlements have different socio-economic characteristics and development controls. The cumulative figure of the residential area, provided More information, crucial to this analysis, such as average plot sizes, by the DDA, does not include qualitative information about the constituent measurements of individual buildings and area occupied by roads and civic housing structures and occupants. However, such information is essential to amenities, is not provided in the DDA plans. We used online mapping tools understand the suitability for solar in different residential colonies/areas. Thus, (Google earth and Google Maps) and referred to the City Map155 for such rather than using the land area figures provided by the DDA, we use a bottom measurements and for detailed analysis of constituent areas. Qualitative up approach in order to analyze this land area type in detail. Using Google information (Such as quality of construction) about areas was obtained through Maps, we first determined the area measurements of individual colonies that site visits and from insights provided by architects and real estate developers. qualified for solar rooftop installations. Colonies that had sufficient solar suitable rooftop space, concrete constructions and occupants that could Our analysis uses the framework used by the DDA Zonal plans, which classifies afford solar installation were qualified for analysis. We then summed up the the entire land area of the city into 9 types: ‘Residential’, ‘Commercial’, individual figures to reach the total land area for the residences. ‘Industrial’, ‘Recreational’, ‘Transport’, ‘Utility’, ‘Government’, ‘Green belt/ Water body’ and ‘Public and Semi-public Facilities’. Our analysis excludes the While pursuing the bottom-up approach, we found that information about land area types – ‘Green belt/water body’ and ‘Recreational’, which include the entire residential area in the city was available in three sections (as regional parks, water bodies and historical monuments. This is because most per the municipality governing the areas) - Residential colonies under the of these areas do not have built structures that can provide roof space. Further, jurisdiction of the Municipal Corporation of Delhi (MCD), residential area under any construction /alteration in such areas would require permissions from the the jurisdiction of the New Delhi Municipal Council (NDMC) and residential Ministry of Culture and the Delhi Municipality. Similarly, the land area type, area maintained by the Delhi Cantonment Board (Delhi Cantt.). The colonies ‘Utility’ that includes areas covered by drains, sanitary landfills, power stations, under the MCD are categorized from A to H, as per the circle rate158 applicable sewage treatment plants and water treatment plants, was excluded from our to each colony. Circle rates are indicative of the economic standing of the analysis. This is because the areas covered by drains and sanitary landfills do colonies, which further provides a sense of the income level of the occupants, not have built structures and are instead occupied by drainage pipes and solid the quality of constructions, the type of housing and the average plot sizes. We waste. Water treatment plants, sewage treatment plants and power stations excluded low income colonies as they are unsuitable for solar installations159. are also largely made up of tanks and processing units offering little or no This categorization provided us with tangible figures and a comprehensive concrete constructions for solar power installations. list of colonies to work with. For residential areas under the NDMC, we have used cumulative figures from the DDA zonal plans. This is because the For all remaining land area types, we first determined the total land area NDMC area largely has government owned colonies that have large rooftops, available for designated uses from the information provided by the Zonal Plans. similar rooftop structures, with no ownership rights for the occupants. In We then discounted for roads, civic amenities, and irrelevant sub categories, if the case of the residential colonies under the Delhi Cantonment area too we any, to reach the total plotted area. This discounting factor for trees roads and have considered all the colonies as they are government owned and have the pedestrian paths ranges from 40-60% for different land area types156. Further, we computed the area of built structures (raw rooftop space) as per the ------157 JJ clusters are households of poorly built congested tenements with inadequate infrastructure specific development controls (coverage rates) pertaining to the type and size and lacking in proper sanitation and drinking water facilities. Unlike slums, these are illegal settlements. ------158 Circle Rate is the minimum rate per sq. meter area on which stamp duty has to be paid by 155 ‘Eicher Delhi City Map’ was used for scaled measurements. the purchaser. This rate is decided by the government authorities for valuation of land in 156 The discounting factor was considered after analyzing sample sites for each land area types and a particular area. after consulting architectural experts. 159 Refer to Table A2 “Assessment of colonies under the jurisdiction of the MCD” for details

76 77 same characteristics as colonies in the NDMC area. For these colonies, we G 100 20 Low Individual houses and Poor have mapped and analyzed all of them using online mapping tools because apartments the area of the cantonment is relatively small and hence measurable. The H 50-100 Oct-20 Low Individual houses and bottom-up approach enabled us to understand the residential area in Delhi apartments with a fair amount of accuracy in order to reach a representative figure for the solar suitable rooftop space available in Delhi’s residences. We have assumed Source: BRIDGE TO INDIA analysis that 20% of the total area can be discounted for having old rooftops, unable to support solar installations. This discount factor is higher than what we have considered for other categories (20% vs. 10%) because we assume that a The total residential area under the jurisdiction of the MCD is classified into 8 larger proportion of residential buildings are on average older than buildings categories, A to H. The basis for such classification is the circle rate applicable in the other categories. After this, the solar suitable rooftop space for Delhi’s to each colony, where colonies under category A have the maximum circle rate residences was assumed to be around 20% of the total rooftop area. This and those under category G have the minimum circle rate. This categorization is because water tanks, ventilating shafts, drying clothes and storage units indicates the economic standings of the colonies. The economic standing occupy most of the area on a residential rooftop leaving limited unoccupied/ further provides a sense of the income level of the occupants, the quality of unobstructed space. constructions, the type of housing and the average size of plots. For example, a colony in category A, such as Sunder Nagar, would commonly have rich Table A1: Solar suitable areas by jurisdictions inhabitants that live in large-sized individual bungalows. We have used this categorization as a framework to assess the suitability of colonies to instate rooftop solar power installations. Residential Total qualified areas as per Total built area/ area after 20% Solar suitable In the analysis, we have considered 10 sample colonies from each category municipalities raw rooftop area discount for old roof top area to determine the income levels of occupants, quality of constructions, type of governing them (km2) rooftops (km2) (km2) housing and average size of pots. After analyzing these factors, we selected categories that indicate the suitability to install solar rooftop PV. Then, using Area under the 79.2 63.32 12.6 Google Maps, we mapped these sample colonies to determine the average MCD area of a colony in a particular category. The average area was further used to Area under the 8.6 6.9 1.4 compute the total raw rooftop area/ built area available for solar in a particular NDMC category. The solar suitable rooftop space was then computed as 20% of the Area under the 5.3 4.3 0.85 total rooftop area. Delhi Cant. TOTAL 93.1 74.5 14.90 Colonies under the categories F , G and H are excluded from our analysis, as they were unsuitable for solar power generation. Using Google Maps and site Source: DDA’s Master Plan 2021, Delhi Zonal Plans, Google Earth and BRIDGE TO INDIA analysis visits, we found that the houses in these colonies have small plot sizes offering a solar suitable rooftop space of 20-30 m2, which barely meets the minimum space160 required for rooftop installation. The poor quality of constructions in such colonies will not be able to endure the weight of rooftop solar Table A2: Assessment of colonies under the jurisdiction of the MCD installations without significant structural improvements. Further, inhabitants Solar of these colonies, belonging to low-income groups, will not be able to afford Average suitable the upfront investment costs associated with solar installations. Some key Colony plot size area Quality of colonies from these categories are Majnu Ka Tila, Shakarpur, Nathu Pura and category (m2) available Income level Principal type of housing Construction Qualification Zakir Nagar. A 900 180 High Individual houses Good Similarly, colonies under category E did not qualify for our analysis, as most B 400 80 Medium to high Individual houses and Good houses in these colonies are low quality constructions. The plot sizes in this apartments category are bigger than those in category F, G and H but are still fairly small C 200-600 40-120 Medium to high Individual houses and Average and may not offer sufficient rooftop space suitable for solar installations. Some apartments key colonies in this category are Paharganj, Yusuf Sarai and Inderpuri. D 250 50 Medium to high Residential complexes, Average apartments and individual Colonies under category D are densely populated with residential complexes houses being the principal type of housing. Some key colonies in this category are E 100-250 20-50 Low to medium Individual houses and Low Mayur Vihar, Paschim Vihar and Rajouri Garden. The residential society as apartments a whole usually owns the rooftop space in such complexes. This may create F 150 30 Low Individual houses and Poor apartments ------160 Minimum unobstructed space required for a 1kW rooftop PV installation is 10 m2 as per ‘RENI- Renewables Insight, Energy Industry Guides’

78 79 problems when obtaining permissions to mount rooftop solar installations m2 and have structures that can endure solar rooftop installations (checked but may also make the solar installations affordable, as costs would be through Google Maps and site visits). In order to determine the residential area divided among a larger group of individuals. The roof space of such residential of the Cantonment area or the area covered by military housing, we mapped complexes is usually occupied by support services such as water tanks, shafts the constituent colonies using Google Maps. We excluded the army hospital and other amenities that are instated for the entire building, leaving less and mess from the mapping exercise to avoid double counting them as they unoccupied space. Despite such issues, with an average plot size of around might be included under the land area type ‘pubic and semi-public facilities’ 250 m2 offering a solar suitable roof space of 50 m2 and concrete building as per the Delhi Master Plan. Manual online mapping provided a more structures, this category qualified for our analysis. The built area/ raw rooftop representative figure and was used it specifically for the Delhi Cantonment, as area available in colonies in category D is around 48.92 km2. its area is relatively small and realistically measureable. Some key colonies inside the Delhi Cantonment area are Chitral lines, Kabul Lines and Colonies under category C, with fairly big plot sizes and good quality Asmara Lines. constructions, are also suitable for rooftop installations. The solar suitable roof space available is around 40 to 120 m2 and occupants largely belong to medium 12.2.2 Commercial buildings income groups. Constituent colonies in this category commonly have a mix of individual houses and apartments/flats. Some key colonies in this category are This land area type includes retail shopping centers, commercial centers, Lajpat Nagar, Punjabi Bagh and Civil Lines. The total built area/rooftop area wholesale markets, warehouses, oil depots and cold storages. Standalone offered by colonies in this category is around 16.87 km2. convenience stores that are integrated into the residential areas are included in the land area type ‘Residential’. Finally, Categories A and B are the most appropriate fit for solar rooftop installations. Colonies in these categories have solar suitable rooftop of up to To determine the solar suitable rooftop space offered by commercial buildings 180 m2, 18 times of what is required. In addition to that, the houses in these we have used cumulative figures provided by the DDA Zonal Plans for reasons colonies are well-constructed individual houses. The wealthiest population similar to that of public and semi-public facilities. The development controls of Delhi inhabits these colonies and is probably in a position to afford such for this land area type are different for different sub heads. Ground coverage installations by itself. Some key colonies in these categories are Friends rate is 25% for district centers, community centers and city centers, 30% for Colony, Sunder Nagar and West End. The total built area/ raw rooftop area of hotels and 40% for local shopping centers as per the development controls colonies in this category is around 13.38 km2. The total built area of residential in the Delhi Master Plan 2021. Average ground coverage rate taken for the colonies under the jurisdiction of the MCD is around 79.2 km2. analysis is 30% as a reasonable number. The total land area designated for commercial use is around 45.59 km2 163 , which offers a built area/rooftop area Residential area under the jurisdiction of the NDMC161 of around 13.6 km2. We assume that 10% of this area belongs to small-sized shops and commercial buildings. Such buildings have an area less than 400 m2 Residential area under the NDMC mainly comprises of either government and offer very little solar suitable roof space (less than 15%). In addition to that, housing areas or affluent residential properties of Delhi. Some key colonies the structures of these buildings are typically weak. These small buildings are under this category are Bapa Nagar, Golf links and Kaka Nagar. Through hence discounted from our analysis, as they will not be able to accommodate Google Maps, we found that the plot sizes within these colonies is 500-1000 the size or bear the weight of the solar installations164. We further discount m2 offering sufficient solar suitable rooftop space. The housing area under 10% for old commercial buildings from the remaining area of commercial the NDMC comprises of well-maintained concrete buildings. All Colonies buildings to get the total qualified rooftop area of 11 km2 for our analysis. in this category are largely similar to colonies in category A of the MCD and are hence considered to have qualified for our analysis. Housing in the area After discounting, commercial consumers are classified into two types: maintained by the NDMC falls in the DDA zone ‘D’. The DDA plans provided the those with an electricity consumption of 100-200 kW that we will call type-1 best numbers to use considering that the area under the NDMC is too large for consumers and those with a consumption greater than 200 kW that we will call actual mapping and the fact that most constituent colonies are similar in terms type-2 consumers. Type-1 and type-2 consumers pay different tariffs, which of size and structure and hence are subject to similar developmental controls. impacts the financial suitability of solar power for them165.

Residential area maintained by the Delhi Cantonment162 Type-1 consumers (medium-sized) account for 75% of the qualified rooftop area i.e. 8.2. km2. They include buildings such as commercial complexes and The Delhi Cantonment houses the Indian army Headquarters, the army shopping areas that have an average area ranging from 2,000-4,000 m2. Type-2 golf course, military housing such as residential complexes and flats, army consumers (large-sized), which account for 25% of the total built area i.e. 2.7 hospitals and public schools. All buildings within this area are well-built km2, comprise of buildings such as shopping malls and hotels that have an structures maintained by the Delhi Cantonment Board with plot sizes of around average size upwards of 4,000 m2. Thus, the total qualified rooftop area for 250-500 m2. We consider all buildings within this area to be qualified to bear commercial consumers is 11 km2. solar installations as they offer solar suitable roof space of around 50-100 ------163 ------As per the Zonal Plans developed by the Delhi Development Authority 161 NDMC or the New Delhi Municipal Council is one of the local governing bodies in Delhi that 164 The 400 m2 cut off for small commercial buildings is double the cut off point for residential administers and maintains around 3% of Delhi’s area. The government of India owns about 80% buildings 200 sq m as rooftops of commercial buildings are typically more cluttered of the buildings under the jurisdiction of the NDMC. (e.g. with cooling towers) than the rooftops of residences. 162 Cantonment was the permanent military station in British India. 165 Refer to chapter 6 “Viability for commercial consumers” for further details

80 81 As per our analysis of commercial buildings using Google Maps, their rooftops Based on our analysis of samples170 through Google Maps and site visits, plot are cluttered with cooling towers, ventilating shafts and water heating systems sizes in industrial areas commonly exceed 400 m2, with roof space occupied by and tanks, reducing the solar suitable roof space. Type-1 consumers have boilers, cooling units and shafts. None the less, there is significant unoccupied/ solar suitable rooftop space that is 30% of their qualified rooftop area unobstructed rooftop space sufficient for solar rooftop installations. (8 km2) giving a total of 2.4 km2. Type-2 consumers have solar suitable rooftop Accommodating for such obstructions, the solar suitable rooftop space for space that is 20% of their qualified rooftop area (2.7 km2) giving a total of both type-1 and type-2 consumers is around 40% of their respective qualified 0.6 km2. Thus, the total solar suitable rooftop space for commercial consumers rooftop area (6.3 km2 each). This gives solar suitable rooftop space of 2.26 is 3 km2. The % of solar suitable space for commercial consumers has been km2 for type-1 and type-2 consumers each, or a total of 4.5 km2 for industrial determined after analyzing sample commercial areas such as Promenade Mall consumers. Some of the key industrial areas in Delhi are Okhla Industrial area, and Nehru place. The solar suitable area is smaller for larger than medium- Badli Industrial area and Wazirpur Industrial area. sized commercial buildings because these larger buildings tend to have an unsuitable architecture or elevated structures ,such as domes and 12.2.4 Government buildings glass panels. This land area type includes areas belonging to the President’s Estate, the 12.2.3 Industrial Buildings Parliament House, government courts (High Court, Supreme Court and the District Courts), integrated government office complexes and land allocated The industrial area of Delhi mainly comprises of textile manufacturing for undetermined government use. In our analysis, we considered unoccupied units, electrical machinery manufacturing units and units that offer repair government land as built-up land based on an understanding that the city of services. This area is characterized by robust concrete buildings and big plot Delhi will eventually be entirely built. Government buildings are commonly sizes, making rooftops well qualified to bear the size and weight of solar concrete structures with average plot sizes upwards of 6,000 m2, which installations. qualifies them for the analysis.

For data related to this land area type, we used cumulative figures provided by According to the DDA Master Plan 2021, all government buildings are subject DDA zonal plans because in industrial areas the building structures and their to similar development controls hence we have used the cumulative land development controls are fairly homogenous. The total land area in Delhi that area figures provided by the DDA plans. The total land area designated for is designated for Industrial use is around 47 km2. 166 Out of this area, 60%167 of government use is around 47 km2 171 and 80% of the area i.e. 37.71 km2 is the area is authorized for Industrial plot development is around 28 km2. The assumed to be plotted for development. The total built area/rooftop area built area/ raw rooftop area available for solar rooftop installations is around available for solar power installations is 30%172 of the plotted land area, 50%168 of the total area available for plot development, which gives around 14 which is 11.3 km2. We further discount 10% of the area for structures that km2. We discount 10% of the area for structures that might be old and unable might be old and unable to bear the weight of solar installations. Thus, the to bear the weight of solar installations. Our analysis further discounts 10% total qualified rooftop is 10 km2. The solar suitable rooftop space offered by for mini-micro and very large industrial buildings (that fall under <10kW and government buildings is assumed to be around 40%, which is fairly larger than >200kW tariff group). Mini-micro industries have been excluded, as they are other land area types, considering their rooftop space is commonly free from too small to accommodate the size of solar installations that could meet their elevator shafts and cooling towers. Structurally these buildings are simple with power needs. The number of large industries on the other hand is very small very little shadowing.This amounts to a solar suitable rooftop space of to 4 km2. and hence such buildings are chosen to be kept out of the analysis. The total qualified rooftop area after discounts is 11.3 km2. 12.2.5 Public and semi-public facilities

Industrial buildings can be categorized into two types, those with 10-100 kW Public and semi-public facilities include hospitals, education/research of electricity consumption that we will call type-1 and those with 100-200 kW universities, colleges, social and cultural complexes, sports complexes, of electricity consumption that we will call type-2. The two types of consumers stadiums, police stations/headquarters, fire stations, disaster management pay different electricity prices which can impact the financial suitability of solar centers, religious buildings, burial grounds and crematoriums. Such buildings power for them169. qualified for our analysis as most facilities included in this category are large, well-built structures. In addition to that, government bodies administer many We roughly assume that 50% of the buildings are type-1 and 50% are type- of these buildings and the government can play a direct role in ensuring that 2 consumers due to lack of information, but assuming that higher number these facilities install rooftop solar installations. of small industrial buildings will offset the smaller number of large-sized industrial buildings. The qualified rooftop area available for type-1 and type-2 We have used the cumulative figures provided by the DDA for this land area consumers is 5.6 km2 for each type. type because there are more than 200 of these facilities, which are very widely distributed across Delhi, and are hence immeasurable. Further, ------166 As per the Zonal plans developed by the Delhi Development Authority 167 As per Industry Zone Guidelines in the Delhi Master Plan 2021 ------168 As per the developmental controls specified in the Master Plan 2021, 50 % is the ground 170 Samples included industrial buildings in Wazirpur, Badli and Naraina industrial area. coverage rate pertains to plot sizes of more than 400 m2. Our analysis assumes that on an 171 As per the Zonal Plans developed by the Delhi Development Authority 2 average, plot sizes in industrial areas exceed 400 m 172 30% ground coverage is permissible for the government integrated office complexes and courts 169 Refer to Chapter 6 “The viability of rooftop solar PV” for details as per the DDA Master Plan 2021

82 83 the development controls for such facilities are almost similar and hence 12.2.7 Factors that could impact the potential for cumulative figures can be used to derive near accurate estimates. rooftop solar

The total land area designated for public and semi-public facilities is around 81 km2 173. Of this, 40% is discounted for trees, roads and other open spaces giving Factors that could increase the Factors that could decrease the a plot areas of 48.81 km2. Of the plots, 30 % is built174 giving a total built area/ potential potential raw rooftop area of 14.64 km2. After discounting for unsuitable public facilities 1.Residential colonies E, F, G and 1. We discount 10-20% for old and such as burial grounds, dairy farms, and religious buildings175, the total built H under the jurisdiction of the MCD structurally weak buildings. If the area for public and semi-public facilities is 13 km2. We further discount 10% have been entirely excluded from percentage of such buildings is of the area for structures that might be old and unable to bear the weight of our analysis. In case these colonies larger, the potential of solar would solar installations (these include police stations and fire stations). This gives a have some solar suitable space, the reduce qualified rooftop potential of around 11.86 km2. potential would increase. 2. In case there are buildings that Public and semi-public facilities can be government owned or privately 2. We have assumed that different have architectural structures owned. Accordingly, they pay different prices for electricity and their viability building types offer different solar (domes, elevations) unfit to bear for solar differs176. We assume177 that 70% of the facilities are government suitable area. For example, average solar installations or that have owned while 30% are private based on the fact that most public and semi- solar suitable space for government building regulations that do not public buildings facilities are government aided. The qualified rooftop space buildings and residential buildings permit solar installations or where is 8 km2 for government owned facilities and 3 km2 for privately owned is assumed to be 40% and 20 % it is not financially viable to build facilities. Based on our analysis through Google Maps and site visits, the solar respectively. If in some cases, structure over and above the existing suitable rooftop space on all public facilities is around 40%. This takes into buildings offer solar suitable spaces roof structure to accommodate consideration that rooftops of most public facilities are occupied by water more than the average assumed then solar installations, the geographic heating systems (mandatory as per government regulations), cooling towers the potential would increase. potential (that does not account for and elevator shafts. Thus, solar suitable rooftop space is a total 4 km2: 3 km2 such factors) for solar in our analysis for government owned and 1 km2 for privately owned facilities. 3. While determining the geographic would decrease. potential, wherever data provided 12.2.6 Transport buildings by the Delhi Master plan was insufficient or unclear, we considered This land area type includes areas covered by airports, bus terminals/depots, the more conservative estimate. For railway stations, metro stations and transport movement infrastructure like example, while mapping the area in train tracks. Train tracks have been excluded from the analysis. Metro stations the Delhi Cantonment for category also fall outside the purview of our analysis as these buildings are mostly ‘Residential’, the Army hospital was made of metallic roof sheeting rather than concrete structures and thus excluded from the mapping exercise. might be unfit to bear the weight of solar installations. Bus depots are large This was done to avoid double open areas for buses to be able to park and do not have any covered, concrete counting as hospitals fall under the structures. Railway stations and airports are the only two sub heads that category ‘public and semi-public qualified for analysis. facilities’ as per the Delhi Master plan. If such areas can be included, Transportation facilities in Delhi that qualify for analysis are airports and the overall potential would increase. railway stations. These are limited in number hence we used online mapping tools to measure their areas. These facilities occupy large stretches of land for 4. We assume that the maximum movement, such as runways, railway platforms and parking, but possess only space required to accommodate a limited built area, which mainly comprises of ticketing and boarding areas. 1kW solar system is 12 m2. If the The total built area/raw rooftop area under this land area type is around 0.16 space occupied by a solar system is km2 that offers a solar suitable roof top space of 0.032 km2. The solar suitable less than this assumption, say 10 m2, roof area is assumed to be 20 % of the total raw rooftop area as most transport the potential will increase. facilities have roof sheeting and the little concrete space left is occupied by support services such as cooling towers, shafts and monitoring units. 5. We discount 10% for smaller shops

------or stand-alone stores that cannot 173 As per the Zonal Plans developed by the Delhi Development Authority accommodate or withstand the size 174 30% coverage rate is applicable to Public and semi public facilities as per and weight of solar installation. In the Delhi Master Plan 2021 case there are some buildings that 175 Discounting factor for irrelevant sub categories is assumed to be 10% as per Bridge to India analysis. These buuldings have been discounted as they have very little buitl space and a solar despite being small are fit to bear installation may not be really wanted by them the weight of the installations, the 176 Refer to Chapter 6 on “The viability of rooftop solar PV” for details potential for solar would increase. 177 Rough assumption is based on the knowledge that within certain categories - e.g. hospitals - there are more government buildings than private buildings (from Eicher Delhi map)

84 85 12.3 wAys to increase Delhi’s Figure A1: Solar PV system geographic potential for solar

Apart from the existing built area in the city, solar installations could be installed over other structures such as drainage pipes or nallahs178. In some other places, land occupied by water bodies/green belts could be utilized by building structures over them that can support solar panels. Also, metro railway tracks are laid on concrete, elevated structures. These structures could be used for solar PV installations.

Building integrated PV (BIPV) systems integrate the photovoltaic modules into the building envelope, including the façade. This system utilizes more area on buildings (not just the roof area/structure available) for solar installations and hence can increase the geographic potential for solar in the city. 12.4 Rooftop solar PV systems

This section deals with the overview of a solar PV system, working of a PV system, different components, their sizing and connections. Further, we have analysed relevance of storage in PV systems, the importance of net metering and the typical PV system sizes across different consumer segments.

Functioning of a PV system

A solar PV power plant converts sunlight into electricity. It does so without any moving parts (unless it has a tracking system) and without generating either noise or pollution. A solar PV system can be installed at any un-shaded location such as on rooftops of buildings, car parking sheds, empty land, or even on top of canals and roads. Typical system sizes range from 240 watts to 100 MW. There is very little difference in the technical design between small kW-sized plants (typically de-centralized, off-grid) and large, MW-sized plants (typically centralized, grid-connected). Solar plants can easily be scaled using independent, modular components such as PV modules, inverters and batteries. The rooftops of buildings would be ideal for the installation of solar PV in Delhi because of the high cost of land in the city. A typical rooftop solar PV system for a household is between 1-10kW. For larger buildings, such as offices or malls, it can reach 100 kW or more. Given Delhi’s average irradiation of 5-5.5 kWh/m2/day, a kW of installed solar PV can generate 4-4.5 kWh of power during daylight hours, equivalent to the amount of power needed for a refrigerator to run for a day.

A solar PV system consists of the following key components 1. Solar PV array (group of modules) 2. Solar inverter 3. Battery 4. Interconnecting devices (junction box, cables, distribution box)

The PV array consists of solar modules interconnected with each other. The modules convert the energy from sunlight, are held on structures made of galvanized iron, mild steel or aluminum and are inclined at a horizontal tilt, facing either south or east-west.

------178 Nallah is an uncovered flowing sewer Source: BRIDGE TO INDIA

86 87 or in case the consumption is not uniformly sufficient throughout the day180, batteries to store energy might be added.

Another factor, which affects the system design, is the timing of electricity consumption. For example, residential consumers in Delhi have a peak demand during the morning (6:00 am – 10:00 am) and evening (6:00 pm – 10:00 pm). These are not peak sunshine hours (10:00 am – 4:00 pm). Residential demand tends to be lower during the day as household members become engaged in daily activities, mostly outside the house (e.g. adults going to work and children going to school). Thus, the peak power production of a PV system does not match the peak demand of residential consumers in Delhi. For industrial and commercial consumers, on the other hand, solar generation coincides more closely with peak demand as most of these sites operate through the day.

Figure A2: Delhi residential demand compared with solar generation181

Source: BRIDGE TO INDIA

The modules are designed to generate current at either 12 or 24 volt. Inverter models can differ in their input voltage requirements in the range of 12 to 1,000 volt. A junction box connects the modules in series or parallel to achieve the optimum voltage required by the inverter.

Solar modules produce direct current (DC). Almost all electrical appliances in India, however, require alternating current (AC) to operate. The function of converting DC to AC is carried out by the inverter. In the case of a battery backup system, the inverter is also connected to the batteries and is Source: Solar irradiation data from NREL responsible for managing the charging and discharging of the batteries. and BRIDGE TO INDIA analysis

The output point of the inverter is connected to the distribution box, which consists of a meter, fuse, a miniature circuit breaker (MCB), and load Solar PV systems could be sized to not exceed the load demand during the connections. Cables connect the solar modules, junction boxes, inverters and day. If they are larger, and solar power is being generated that exceeds distribution boxes. consumption at that point in time, wastage can be avoided by storing the excess power. Alternatively, excess power could be injected into the grid. In this The capacity of the solar PV system depends on the amount of electricity (kWh) case, metering would be required to measure energy transactions between the required per day by a consumer and the shadow and obstruction free space PV system and the grid. As shown in the graph above, 40% solar generation available on the rooftop. For example, a 2 kWp load operating for 10 hours will be unused if there is no net metering. requires a PV system of 5 kWp179 . Further, 1 kWp of solar PV requires 10 m2 of shadow free area. Therefore, a 5 kWp system would require 50 m2. In addition, if the consumption occurs during non-sunshine hours (6:00 pm to 6:00 am) ------180 If loads have no fixed time of operation or might be operational over a couple of hours. For example there can be a 10 kW load being powered for just 2 hours a day. This would require a system of 5 kWp. A 5 kWp system cannot generate enough electricity to power a 10 kWp load at ------any time of the day; Hence we require batteries that store energy 179 The size of a solar PV system is usually higher than the operational load as solar has an average 181 Energy consumption data mapped - based on a residential house with a load of connected load of Capacity Utilization Factor (CUF) of 18% as compared to 80% for coal power plants. 1 kW of solar 10 kW and operating load of maximum 5 kW. Solar irradiation data of 02.08.2011, source NREL. gives 4 kWhs at a CUF of 18%. So, For 20 kWhs of power, we would need a system of 5 kW. Solar system size - 5 kWp

88 89 Rooftop solar PV with storage The CEA has initiated steps to set standards and guidelines for the integration of solar PV systems in to the grid. A report on grid connectivity of solar PV Storage in solar PV systems is required to provide stable backup power when is under formulation at the CEA. A draft is to be shared with the public for the solar energy is not available (at night) or not adequate to meet the entire comments by end-June. The CEA’s move is based on its acknowledgement load demand. Solar energy is an intermittent source of power. The power that decentralized solar PV can play a key role in bridging the country’s energy generation can vary with a change in sunshine due to, for example, a sudden deficit and is set to take off now. During our interactions with senior officials cloud cover. Batteries can be used to store solar power to safeguard against a at the CEA, we were told that “solar PV is the future for this country, and we short-term fall in solar power generation. Intermittency can also be avoided have to make sure that there are standards and guidelines in place to support by connected the solar PV system to the grid. In this case the grid provides the its integration with the grid”. Various metering arrangements covering grid extra energy at times of inadequate sunshine. interaction of a PV system with battery, without battery, with different load battery back-ups, with different load DG back-ups, and with DG and battery Another application of storage is to protect against grid outages. During an back-up combinations, have been laid down in the draft report. The DERC184 outage it is possible that solar generation is inadequate to meet the load though has the final authority to determine the metering arrangements that demand (e.g. if it occurs outside sunshine hours). In such a case, the stored would be applicable in the city. energy can be utilized to provide a stable output of power. If the grid condition is good and power outages are rare, batteries would probably be avoided as Rooftop solar PV system sizes they add significantly to the system cost, adding 25%.182 Batteries also need to be replaced every three to five years.183 Since Delhi does not experience long Commercial consumers in Delhi (type-1, and type-2 and privately owned public power cuts, batteries need not be an essential part of the PV system. Storage and semi-public facilities) such as malls and office complexes have an average might, however, be an attractive option for Delhi’s DISCOMs as it can be utilized load requirement above 1 MW. But due to the presence of cooling towers, air to offset expensive peak load power. conditioning units, ventilating shafts, boilers and tanks, limited rooftop space is available for solar. As per our analysis on rooftop potential in Chapter 1 of Rooftop solar PV with net metering this report, type-1 consumers accounting for 75% of the total built commercial area include buildings such as small office complexes have an average area There are two common ways in which owners of kW-scale rooftop solar PV of 400-2,000 m2. These buildings offer a solar suitable roof space of 30%, plants can be compensated for feeding electricity into the grid: FiTs and net which is upwards of 120 m2. This space can typically accommodate PV systems metering. For FiTs, solar power generated and fed into the grid is measured of sizes exceeding 10 kWp185. Given the range, we assume 25 kWp to be the through a separate meter and then given a price (the FiT) through which the standard system size for type-1 commercial consumers in our analyses. Type owner is compensated for the electricity generated. The advantage is, that 2 consumers, account for 25% of the total built commercial area comprising the price for solar power and the amount of solar power generated can be of buildings that have an average size ranging from 2,000 m2 to 12,000 m2. The determined independently. This method is useful where either the cost of solar solar suitable space offered by these buildings is 20% that amounts to solar power far exceeds the cost of grid power and/or where the generation of solar suitable space ranging from 400 to 2,400 m2. Type 2 consumers also include power far exceeds the on-site consumption needs. A risk is the potential for privately owned public and semi-public facilities that account for 30% of the fraud through channeling non-solar power through the solar meter and thus built area of public and semi-public facilities. comprise of private hospitals and inflating the amount of power for which the – usually high – solar FiT is paid. A educational institutes. These buildings too typically have rooftop sizes 2,000- household-level FiT is offered in, for example, Germany. 12,000 m2 and offer solar suitable space of 35% or between of 700-4,200 m2. Type-2 buildings can thus typically accommodate a PV system ranging from Under net metering, on the other hand, conventional electricity and solar 33-350 kWp186. Given the range, we assume 50 kWp to be the standard system electricity are traded at the same tariff. The billing in this case is based on size for type-2 consumers in our analyses. the net energy imported (energy consumed minus energy generated and fed into the grid). In case more energy is generated than consumed, the utility can Industrial consumers have fewer obstructions as compared to commercial adjust the excess in a future billing period (this would be akin to “banking” the rooftops. As per our analysis in Chapter 1 of this report, industrial consumers power), rather than giving a monetary compensation, as in the case of FiTs. have 40% of roof space available for solar. The solar suitable space for type- However, over the long term, the amount of solar power that can be generated 1 consumers ranges from 120-320 m2 which can typically accommodate PV and monetized through net metering will be limited by the amount of power systems of sizes ranging from 10-27 kWp187. For type-1 consumers, we assume consumed, where, at most, the consumer can feed as much power back into a standard system size of 20 kWp for our analyses.Type-2 consumers are the grid as he draws from the grid so that the electricity bill is “zero”. Net bigger industrial buildings that have an average area ranging from 800-1,500 metering is popular in, for example, the USA and Japan. m2. The solar suitable area of these buildings ranges from 320-600 m2 that can typically accommodate PV system of sizes ranging from 27-50 kWp188. For type-2 consumers, we assume a standard system size of 40 kWp.

------184 DERC regulates the operation of the power system and sets standards for the electricity industry 182 5 kWp system with battery costs M750,000 (€115,384, $150,000) as compared to M600,000 (standards related to quality, continuity and reliability of service) in the NCT of Delhi. (€92,307 $120,000) for a non battery backup system. Storage has been calculated for 2.5 kW load 185 Based on Bridge to India’s estimate that 12 m2 of rooftop space can accommodate a 1kW for 7 hours. solar power installation 183 Refer to the section on “The Viability of Rooftop Solar PV” for further details 186 Ibid. 187 Ibid. 188 Ibid.

90 91 Residential consumers of have rooftop areas between 200-1,000 m2. With only losses or damages. This injection of DC power into the grid can be avoided 20% of their rooftops suitable for solar, the have a space availability between by using an isolation transformer at the output of the inverter. An isolation 40-200 m2.This gives them a potential between 3-17 kWp189. Given the range, transformer would block transmission of DC signals from one circuit to the for residential consumers we assume a standard system size of 5 kWp for other, but allow AC signals to pass. Coupled with an isolation transformer, our analyses. inverters are able to feed electricity into the grid with a maximum permissible DC component of 1% in the AC current. Government buildings such as courts and office complexes and government owned public and semi-public facilities have rooftop areas of 1,000-13,000 m2. Electrical disturbance caused by non-linear loads The solar suitable space available on government buildings is assumed to be 40% of the built area, while that on government owned public and semi-public Electrical disturbance (the permanent modification of the voltage and current facilities is 35%. This translates to solar suitable space between 350-5,200 m2, sinusoidal wave shapes), generating harmonics, is created by the presence of which can accommodate PV systems of 29-433 kWp190. However, government non-linear192 components in an electrical system. Harmonics in the power grid buildings are usually in close proximity of each other, allowing for the bundling can overload equipment, interfere with telephone circuits and broadcasting and of PV systems. Bundling to create projects of a minimum 2,000 kW can reduce lead to metering errors. Typically, total voltage harmonic distortion, individual costs significantly. Thus, we assume standard system sizes of 2,000 kWp for harmonic distortion and total current harmonic distortion produced by a PV government consumers. system should not be allowed to exceed 5%, 3% and 8% respectively, such that electrical disturbances do not affect the quality of power in the grid. 12.5 Solar PV grid connectivity challenges and solutions Unintentional islanding

Voltage fluctuation and imbalance Unintentional islanding occurs when a solar PV system continues to supply electricity to the grid in the event of a larger grid failure. This can cause safety The solar energy generated by a PV system is dependent on the availability of issues for technicians and the general public as a power line that might be sunlight. Power generated can vary drastically throughout the day, sometimes considered off is, in fact, still powered. Unintentional islanding is a well-known within seconds, because of, for example, cloud-cover and weather changes. problem and, therefore, most inverters now possess anti-islanding features This causes rapid voltage fluctuations that can hamper devices linked to the whereby PV systems are automatically disconnected from the grid in the event transmission network and can, in some cases, overheat and even melt the of a power outage. In addition to that, a manual disconnect switch could also power lines. Inverters can be designed to regulate PV system voltage and be provided in the PV system to isolate the grid connection. Despite these provide for communication between the grid operator and the PV inverters. measures, unintentional islanding remains a particular concern when PV and Such communication can allow for improved control of voltage fluctuations in other systems without anti-islanding features, such as diesel generators, are the entire grid. connected to the same line section. These machines may mimic normal grid conditions, causing the PV inverters to stay online and create unintentional Further, voltage fluctuations can take place due to the improper functioning of islanding. At the moment, there are no finalized solutions to this problem. an inverter and can be problematic if these fluctuations move outside specified values. Excessive under-voltage can cause “brownouts” characterized by the Reverse power flow and voltage fluctuations dimming of lights and inability to power some equipment. Excessive over- voltage can damage and decrease the life of electronic equipment. To avoid High levels of PV penetration may cause reverse power flow. Reverse power such scenarios, voltage fluctuations need to stay within specific limits191, flow occurs when in the case of weak or/and long networks, voltage rises beyond which the PV system is required to automatically disconnect from when consumption at the consumer’s end is low and the power fed into the the grid. grid by the PV system is high. In such a case, voltage increase may cause the electricity to change direction and flow through the transformer to the higher Transmission of unwanted current into the grid voltage. This is can lead to heating up of the distribution transformer and transmission lines. Another problem can arise when the power generated by The current in the grid which is supplied for end-use is AC. The current the large number PV systems reduces at once due to the inherent intermittent generated by the solar PV system is DC. This is stored into the battery system nature of solar. This sudden reduction in power can destabilize the grid. (if there is one), also in DC, and then converted into AC by the PV inverter, which in turn is either fed into the grid or consumed. There is a possibility that the PV inverter passes unwanted DC current into the AC driven network of the grid, which can lead to overheating of distribution power transformers, power

------189 190 Ibid. Ibid. 192 Linear loads are electrical loads where the voltage and current waveforms are sinusoidal. The 191 For step changes (which may occur repetitively) the limit stated by CEA is 1.5 %. For occasional current at any time is proportional to the voltage. Non-linear loads are electrical loads where fluctuations other than step changes the maximum permissible limit is 3%. Further, the voltage current is not proportional to voltage. The nature of non-linear loads is to generate “harmonics” imbalance or maximum deviation from phase voltage at 33kV and is not allowed to exceed 3% or distortion in the current waveform. Non-linear loads include computers or refrigerators.

92 93 13 glossary of Terms

1BOG One block off the grid AC Alternating Current BOS Balance-of-system BRPL BSES Rajdhani Power Limited BYPL BSES Yamuna Power Limited CEA Central Electricity Authority CERC Central Electricity Regulatory Commission CNG Compressed Natural Gas CSR Corporate Social Responsibility CUF Capacity Utilization Factor CUNY City University of New York DC Direct Current DDA Delhi Development Authority DERC Delhi Electricity Regulatory Commission DISCOM Distribution Company DOE Department of Energy EPC Engineering, Procurement and Construction FiT Feed-in-Tariffs GBI Generation Based Incentive GIS Geographical Information System GPCL Gujarat Power Corporation Limited IIT Indian Institute of Technology IRR Internal Rate of Return JJ Jhugi Jhompri Clusters LCOE levelized Cost of Energy MCB Miniature Circuit Breaker MCD Municipal Corporation of Delhi MNRE Ministry of New and Renewable Energy NCT National Capital Territory NDMC New Delhi Municipal Council NOx Nitrogen Oxide NREL National Renewable Energy Laboratory NSM Jawaharlal Nehru National Solar Mission PPA Power Purchase Agreement PPP Private Public Partnership PV Photovoltaic REC Renewable Energy Certificates RESCO Renewable Energy Service Company RPO Renewable Purchase Obligations RPS Renewable Portfolio Standard RSPM Respirable Suspended Particulate Matter TPDDL Tata Power Delhi Distribution Limited VAT Value Added Tax WHO World Health Organization

94 95 BRIDGE TO INDIA is a consulting company focusing Greenpeace is a global organization that uses on the solar market in India. The company was non-violent direct action to tackle the most critical founded in 2008 and is based in New Delhi, Bangalore, threats to our planet‘s biodiversity and environment. Munich and Hamburg. BRIDGE TO INDIA offers solar Greenpeace is a non-profit organization, present market Intelligence, strategic consulting and project in 40 countries across Europe, the Americas, development services to investors, companies and Africa, Asia and the Pacific. It speaks for 2.8 million institutions. As part of market intelligence, we provide supporters worldwide, and inspires many millions comprehensive, analytical and up-to-date research more to take action every day. To maintain its on the Indian solar market through various reports. independence, Greenpeace does not accept donations Our strategic consulting expertise lies in assisting from governments or corporations but relies on large Indian and international clients to engage contributions of individual supporters and foundation the solar market in India. We also offer solar PV grants. Greenpeace has been campaigning against project development services with a strong focus on environmental degradation since 1971 when a commercially attractive rooftop business models. small boat of volunteers and journalists sailed into Amchitka, an area west of Alaska, where the US At BRIDGE TO INDIA, we seek to provide innovative Government was conducting underground nuclear and business-driven solutions for our clients. We do tests. This tradition of ‘bearing witness’ in a non- so by combining strong subject specific knowledge violent manner continues today, and ships are an in our core areas related to solar energy with an important part of all its campaign work. interdisciplinary approach, bringing together the financial, technical, socio-economic, regulatory and Contact entrepreneurial aspects of business. [email protected]

Contact 60, Wellington Street [email protected] Richmond Town Bangalore 560025 BRIDGE TO INDIA Pvt. Ltd. N-117, Panchsheel Park New Delhi 110017 India